exercise.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# EGamma short exercise\n",
    "This is the companion notebook to [SWGuideCMSDataAnalysisSchoolLPC2025ShortExEGamma](https://twiki.cern.ch/twiki/bin/view/CMS/SWGuideCMSDataAnalysisSchoolLPC2025ShortExEGamma).\n",
    "\n",
    "Photons and electrons play a crucial role in the success of CMS. They were vital in the discovery of the Higgs boson and play an important role in searches for new physics and standard model measurements. The CMS detector features a high-resolution electromagnetic calorimeter (ECAL) and the world's largest active silicon tracker. These detectors provide clean object identification and excellent energy and momentum resolution of photons and electrons. Within CMS, the [EGamma POG](https://twiki.cern.ch/twiki/bin/view/CMS/EgammaPOG) has the responsibility to \"study, develop, characterize and validate the tools to identify and reconstruct electrons and photons.\" In this exercise, we will examine photon and electron identification and study how these objects are used in physics analyses. We will start with [NanoAOD](https://twiki.cern.ch/twiki/bin/view/CMSPublic/WorkBookNanoAOD) samples.\n",
    "\n",
    "Table of contents:\n",
    "   * [Finding samples](#Finding-samples)\n",
    "   * [Using NanoAOD](#Using-NanoAOD)\n",
    "   * [Array programming primer](#Array-programming-primer)\n",
    "   * [Ground truth](#Ground-truth)\n",
    "   * [Discriminating variables](#Discriminating-variables)\n",
    "      * [Shower shape](#Shower-shape)\n",
    "      * [Isolation](#Isolation)\n",
    "      * [Electron-specific variables](#Electron-specific-variables)\n",
    "      * [Photon-specific variables](#Photon-specific-variables)\n",
    "   * [Cut optimization](#Cut-optimization)\n",
    "   * [Tag and probe](#Tag-and-probe)\n",
    "   * [Scale Factors](#Scale-Factors)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Finding samples\n",
    "Let's find two simulation samples to use for the remainder of the exercise: a $Z\\rightarrow \\ell^{+}\\ell^{-}$ sample, which will contain two electrons in slightly more than 1/3 of the events (**Q: why greater than a third?**); and a sample of $\\gamma+\\text{jets}$ events, where a photon is radiated off either initial- or final-state quarks.  To find samples via a web browser, one can go to the [DAS client](https://cmsweb.cern.ch/das/) page.  Here, we use the `dasgoclient` command-line utility, using Jupyter notebook's shell execution shortcut (the ! symbol).\n",
    "\n",
    "The query format is `dasgoclient --query=\"key requirement [requirement ...]\"` where the key is what type of record we are looking for, and each requirement helps filter results.  Here we want to find some datasets that match a wildcard pattern.  Datasets in CMS (mostly) follow a structured naming convention,\n",
    "\n",
    "```/Primary/Campaign-Conditions-Revision/Datatier```,\n",
    "\n",
    "where: `Primary` is the primary dataset type, either a particular set of triggers for real data, or a particular physics process for simulation; `Campaign` encodes the data reprocessing or simulation generating campaign, for which the most current options for Ultra Legacy campaigns are [listed here](https://twiki.cern.ch/twiki/bin/view/CMS/PdmVRun2LegacyAnalysis); `Conditions` specifies in more detail the reconstruction and/or simulation conditions of the dataset; `Revision` allows for replacement of invalid datasets or extensions (e.g. `_ext1` you might sometimes see); and `Datatier` describes what kind of format and content is in the ROOT files inside the dataset.  Below, we search for a simulation dataset for the Drell-Yan ($Z/\\gamma^{*}$) process, where the decay is to two leptons (`ToLL`) and a generator-level mass cut of 50 GeV is imposed.  Adding `_TuneCP5` helps reduce the number of matches where additional generator cuts are present or alternative shower tunes are being used (you can learn more about shower tunes in the generators exercise).  Try removing it to see what kind of exclusive phase spases samples are commonly generated for.  For the generation campaign, we choose the Autumn18 production, which corresponds to the 2018 data-taking conditions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/DYJetsToLL_M-50_TuneCP5_13TeV-amcatnloFXFX-pythia8/RunIISummer20UL18NanoAODv9-106X_upgrade2018_realistic_v16_L1v1-v2/NANOAODSIM\r\n",
      "/DYJetsToLL_M-50_TuneCP5_13TeV-madgraphMLM-pythia8/RunIISummer20UL18NanoAODv9-106X_upgrade2018_realistic_v16_L1v1-v1/NANOAODSIM\r\n",
      "/DYJetsToLL_M-50_TuneCP5_13TeV-madgraphMLM-pythia8/RunIISummer20UL18NanoAODv9-106X_upgrade2018_realistic_v16_L1v1_ext1-v1/NANOAODSIM\r\n"
     ]
    }
   ],
   "source": [
    "!dasgoclient --query=\"dataset dataset=/DYJetsToLL_M-50_TuneCP5*/RunIISummer20UL18NanoAODv9-106X*/NANOAODSIM\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We see three results, all are produced with the [MadGraph5_aMC@NLO](https://twiki.cern.ch/twiki/bin/view/CMS/CitationsForGenerators#MadGraph5_a_MCatNLO) generator, where one simulates the Z+jets process at leading order (`madgraphMLM`), one at NLO (`amcatnloFXFX`), and one which is just an extension to increase the number of simulated events.  From here, it is possible albeit convoluted to see the original run cards that configured the MadGraph generator.  We tracked them down for you, so if you are curious you can see what actually configures the event generation: [madgraphMLM card](https://github.com/cms-sw/genproductions/tree/1e06dc8f0c86ab6fdb547b27c43c92b022608b80/bin/MadGraph5_aMCatNLO/cards/production/2017/13TeV/DYJets_HT_LO_MLM/DYJets_HT_mll50/DYJets_HT-incl), [amcatnloFXFX card](https://github.com/cms-sw/genproductions/tree/59237cf5f41aed600b63fa8d071441129aeb6f9a/bin/MadGraph5_aMCatNLO/cards/production/2017/13TeV/dyellell012j_5f_NLO_FXFX\n",
    ").\n",
    "\n",
    "Let's use the LO sample.  Below, we query the files belonging to the dataset, and save them to a python variable."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of files: 71\n",
      "First file name: /store/mc/RunIISummer20UL18NanoAODv9/DYJetsToLL_M-50_TuneCP5_13TeV-madgraphMLM-pythia8/NANOAODSIM/106X_upgrade2018_realistic_v16_L1v1-v1/280000/525CD279-3344-6043-98B9-2EA8A96623E4.root\n"
     ]
    }
   ],
   "source": [
    "dy_files = !dasgoclient --query=\"file dataset=/DYJetsToLL_M-50_TuneCP5_13TeV-madgraphMLM-pythia8/RunIISummer20UL18NanoAODv9-106X_upgrade2018_realistic_v16_L1v1-v1/NANOAODSIM\"\n",
    "print(\"Number of files:\", len(dy_files))\n",
    "print(\"First file name:\",dy_files[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, for the photon event samples, since the inclusive production phase space is so large, we have to impose a minimum transverse energy sum requirement (`HT`, i.e. $\\sum_{\\text{particles}} p_{T}$) to be able to generate enough events in the phase space we are interested in (which is dictated by photon trigger thresholds)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/GJets_HT-200To400_TuneCP5_13TeV-madgraphMLM-pythia8/RunIISummer20UL18NanoAODv9-106X_upgrade2018_realistic_v16_L1v1-v1/NANOAODSIM\r\n",
      "/GJets_HT-400To600_TuneCP5_13TeV-madgraphMLM-pythia8/RunIISummer20UL18NanoAODv9-106X_upgrade2018_realistic_v16_L1v1-v1/NANOAODSIM\r\n",
      "/GJets_HT-40To100_TuneCP5_13TeV-madgraphMLM-pythia8/RunIISummer20UL18NanoAODv9-106X_upgrade2018_realistic_v16_L1v1-v1/NANOAODSIM\r\n",
      "/GJets_HT-600ToInf_TuneCP5_13TeV-madgraphMLM-pythia8/RunIISummer20UL18NanoAODv9-106X_upgrade2018_realistic_v16_L1v1-v1/NANOAODSIM\r\n"
     ]
    }
   ],
   "source": [
    "!dasgoclient --query=\"dataset dataset=/GJets_HT-*/RunIISummer20UL18NanoAODv9-106X*/NANOAODSIM\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's use just one HT bin for our photon sample.  We will investigate how much that restricts the generated photon $p_{T}$."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of files: 50\n"
     ]
    }
   ],
   "source": [
    "gjets_files = !dasgoclient --query=\"file dataset=/GJets_HT-200To400_TuneCP5_13TeV-madgraphMLM-pythia8/RunIISummer20UL18NanoAODv9-106X_upgrade2018_realistic_v16_L1v1-v1/NANOAODSIM\"\n",
    "print(\"Number of files:\", len(gjets_files))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The files we've identified above are all LFNs (Logical File Names), and they alone don't specify where and how to read them.  For that, we need to specify a PFN (Physical File Name).  CMS operates a distributed data access system known as [AAA](https://twiki.cern.ch/twiki/bin/view/CMSPublic/WorkBookXrootdService) which makes files available via the [xrootd protocol](https://xrootd.slac.stanford.edu/) regardless of their storage location.  In short, adding the prefix `root://cmsxrootd.fnal.gov/` to all the files makes them readable."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "root://cmsxrootd.fnal.gov//store/mc/RunIISummer20UL18NanoAODv9/DYJetsToLL_M-50_TuneCP5_13TeV-madgraphMLM-pythia8/NANOAODSIM/106X_upgrade2018_realistic_v16_L1v1-v1/280000/525CD279-3344-6043-98B9-2EA8A96623E4.root\n",
      "root://cmsxrootd.fnal.gov//store/mc/RunIISummer20UL18NanoAODv9/GJets_HT-200To400_TuneCP5_13TeV-madgraphMLM-pythia8/NANOAODSIM/106X_upgrade2018_realistic_v16_L1v1-v1/2560000/8316E3E3-C9FD-914C-9AE2-DF47143B5BD5.root\n"
     ]
    }
   ],
   "source": [
    "redirector = 'root://cmsxrootd.fnal.gov/'\n",
    "dy_files = [redirector + f for f in dy_files]\n",
    "gjets_files = [redirector + f for f in gjets_files]\n",
    "\n",
    "print(dy_files[0])\n",
    "print(gjets_files[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Using NanoAOD\n",
    "In previous iterations of this exercise, students started by running an _NTuplizer_ over MiniAOD samples.  The MiniAOD datatier contains enough detailed information to cover essentially all typical CMS analyses.  The EGamma-related Ultra Legacy content and recipes are discussed further in [EgammaUL2016To2018](https://twiki.cern.ch/twiki/bin/view/CMS/EgammaUL2016To2018) TWiki. \n",
    "\n",
    "A plethora of nTuplizers have been written over the past few years, all essentially extracting similar information from MiniAOD.  In an attempt to reduce code maintenence burden, a common nTuple, [NanoAOD](https://twiki.cern.ch/twiki/bin/view/CMSPublic/WorkBookNanoAOD) was created.  It contains less information than MiniAOD, and is a much more compact file format, averaging 1-2 kB per event vs. MiniAOD's 30-50 kB per event.  In addition, NanoAOD saves all information in basic types (like `int`, `float`, etc.) rather than the [EDM format](https://twiki.cern.ch/twiki/bin/view/CMSPublic/WorkBookCMSSWFramework#EdM) used for all parent datatiers.  This means that no CMSSW environment is necessary to read these files.  In fact, we will not even use the ROOT framework to analyze these files in this exercise!\n",
    "\n",
    "For analyses that require information not available in standard NanoAOD, one can create _private NanoAOD_, where  missing variables and/or CMSSW framework modules are added to the [configuration](https://github.com/cms-sw/cmssw/blob/master/PhysicsTools/NanoAOD/python/nano_cff.py) and then CRAB workflows are used to re-extract the information from the parent MiniAOD dataset.  The CMS [cross-POG forum](https://indico.cern.ch/category/5517/) exists to also allow users to petition for certain variables to be added to central NanoAOD.  For the purposes of this exercise, all standard EGamma variables can already be found in NanoAOD.\n",
    "\n",
    "Below, we examine the content of a NanoAOD file using [uproot](https://uproot.readthedocs.io/en/latest/?badge=latest#)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'tag;1': <TObjString 'untagged' at 0x7f2d208586d0>,\n",
       " 'Events;1': <TTree 'Events' (1628 branches) at 0x7f2d20862340>,\n",
       " 'LuminosityBlocks;1': <TTree 'LuminosityBlocks' (2 branches) at 0x7f2d20862880>,\n",
       " 'Runs;1': <TTree 'Runs' (8 branches) at 0x7f2d208628b0>,\n",
       " 'MetaData;1': <TTree 'MetaData' (1 branches) at 0x7f2d208627c0>,\n",
       " 'ParameterSets;1': <TTree 'ParameterSets' (1 branches) at 0x7f2cb7a75f70>}"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import uproot\n",
    "file = uproot.open(dy_files[0])\n",
    "dict(file)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The item of interest here is the `Events` [TTree](https://root.cern.ch/doc/master/classTTree.html) object.  Let's see what is inside it:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "name                 | typename                 | interpretation                \n",
      "---------------------+--------------------------+-------------------------------\n",
      "run                  | uint32_t                 | AsDtype('>u4')\n",
      "luminosityBlock      | uint32_t                 | AsDtype('>u4')\n",
      "event                | uint64_t                 | AsDtype('>u8')\n",
      "HTXS_Higgs_pt        | float                    | AsDtype('>f4')\n",
      "HTXS_Higgs_y         | float                    | AsDtype('>f4')\n",
      "HTXS_stage1_1_cat... | int32_t                  | AsDtype('>i4')\n",
      "HTXS_stage1_1_cat... | int32_t                  | AsDtype('>i4')\n",
      "HTXS_stage1_1_fin... | int32_t                  | AsDtype('>i4')\n",
      "HTXS_stage1_1_fin... | int32_t                  | AsDtype('>i4')\n",
      "HTXS_stage1_2_cat... | int32_t                  | AsDtype('>i4')\n",
      "HTXS_stage1_2_cat... | int32_t                  | AsDtype('>i4')\n",
      "HTXS_stage1_2_fin... | int32_t                  | AsDtype('>i4')\n",
      "HTXS_stage1_2_fin... | int32_t                  | AsDtype('>i4')\n",
      "HTXS_stage_0         | int32_t                  | AsDtype('>i4')\n",
      "HTXS_stage_1_pTjet25 | int32_t                  | AsDtype('>i4')\n",
      "HTXS_stage_1_pTjet30 | int32_t                  | AsDtype('>i4')\n",
      "HTXS_njets25         | uint8_t                  | AsDtype('uint8')\n",
      "HTXS_njets30         | uint8_t                  | AsDtype('uint8')\n",
      "nboostedTau          | uint32_t                 | AsDtype('>u4')\n",
      "boostedTau_charge... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "boostedTau_eta       | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "boostedTau_leadTk... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "boostedTau_leadTk... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "boostedTau_leadTk... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "boostedTau_mass      | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "boostedTau_neutra... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "boostedTau_phi       | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "boostedTau_photon... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "boostedTau_pt        | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "boostedTau_puCorr    | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "boostedTau_rawAnt... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "boostedTau_rawIso    | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "boostedTau_rawIso... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "boostedTau_rawMVA... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "boostedTau_rawMVA... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "boostedTau_rawMVA... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "boostedTau_charge    | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "boostedTau_decayMode | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "boostedTau_jetIdx    | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "boostedTau_rawAnt... | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "boostedTau_idAnti... | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "boostedTau_idAntiMu  | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "boostedTau_idMVAn... | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "boostedTau_idMVAo... | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "boostedTau_idMVAo... | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "btagWeight_CSVV2     | float                    | AsDtype('>f4')\n",
      "btagWeight_DeepCSVB  | float                    | AsDtype('>f4')\n",
      "CaloMET_phi          | float                    | AsDtype('>f4')\n",
      "CaloMET_pt           | float                    | AsDtype('>f4')\n",
      "CaloMET_sumEt        | float                    | AsDtype('>f4')\n",
      "ChsMET_phi           | float                    | AsDtype('>f4')\n",
      "ChsMET_pt            | float                    | AsDtype('>f4')\n",
      "ChsMET_sumEt         | float                    | AsDtype('>f4')\n",
      "nCorrT1METJet        | uint32_t                 | AsDtype('>u4')\n",
      "CorrT1METJet_area    | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "CorrT1METJet_eta     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "CorrT1METJet_muon... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "CorrT1METJet_phi     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "CorrT1METJet_rawPt   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "DeepMETResolution... | float                    | AsDtype('>f4')\n",
      "DeepMETResolution... | float                    | AsDtype('>f4')\n",
      "DeepMETResponseTu... | float                    | AsDtype('>f4')\n",
      "DeepMETResponseTu... | float                    | AsDtype('>f4')\n",
      "nElectron            | uint32_t                 | AsDtype('>u4')\n",
      "Electron_dEscaleDown | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_dEscaleUp   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_dEsigmaDown | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_dEsigmaUp   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_deltaEtaSC  | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_dr03Ecal... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_dr03Hcal... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_dr03TkSumPt | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_dr03TkSu... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_dxy         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_dxyErr      | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_dz          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_dzErr       | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_eCorr       | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_eInvMinu... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_energyErr   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_eta         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_hoe         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_ip3d        | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_jetPtRelv2  | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_jetRelIso   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_mass        | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_miniPFRe... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_miniPFRe... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_mvaFall1... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_mvaFall1... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_pfRelIso... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_pfRelIso... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_phi         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_pt          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_r9          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_scEtOverPt  | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_sieie       | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_sip3d       | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_mvaTTH      | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Electron_charge      | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Electron_cutBased    | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Electron_jetIdx      | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Electron_pdgId       | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Electron_photonIdx   | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Electron_tightCharge | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Electron_vidNeste... | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Electron_vidNeste... | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Electron_convVeto    | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Electron_cutBased... | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Electron_isPFcand    | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Electron_jetNDauC... | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Electron_lostHits    | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Electron_mvaFall1... | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Electron_mvaFall1... | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Electron_mvaFall1... | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Electron_mvaFall1... | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Electron_mvaFall1... | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Electron_mvaFall1... | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Electron_seedGain    | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "nFatJet              | uint32_t                 | AsDtype('>u4')\n",
      "FatJet_area          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_btagCSVV2     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_btagDDBvLV2   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_btagDDCvBV2   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_btagDDCvLV2   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_btagDeepB     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_btagHbb       | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_deepTagMD_... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_deepTagMD_... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_deepTagMD_... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_deepTagMD_... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_deepTagMD_... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_deepTagMD_... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_deepTagMD_... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_deepTagMD_... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_deepTagMD_... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_deepTagMD_... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_deepTag_H     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_deepTag_QCD   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_deepTag_QC... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_deepTag_Tv... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_deepTag_Wv... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_deepTag_Zv... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_eta           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_mass          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_msoftdrop     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_n2b1          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_n3b1          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_particleNe... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_particleNe... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_particleNe... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_particleNe... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_particleNe... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_particleNe... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_particleNe... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_particleNe... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_particleNe... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_particleNe... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_particleNe... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_particleNe... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_phi           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_pt            | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_rawFactor     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_tau1          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_tau2          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_tau3          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_tau4          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_lsf3          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FatJet_jetId         | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "FatJet_subJetIdx1    | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "FatJet_subJetIdx2    | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "FatJet_electronId... | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "FatJet_muonIdx3SJ    | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "FatJet_nConstituents | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "nFsrPhoton           | uint32_t                 | AsDtype('>u4')\n",
      "FsrPhoton_dROverEt2  | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FsrPhoton_eta        | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FsrPhoton_phi        | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FsrPhoton_pt         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FsrPhoton_relIso03   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "FsrPhoton_muonIdx    | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "nGenJetAK8           | uint32_t                 | AsDtype('>u4')\n",
      "GenJetAK8_eta        | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenJetAK8_mass       | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenJetAK8_phi        | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenJetAK8_pt         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "nGenJet              | uint32_t                 | AsDtype('>u4')\n",
      "GenJet_eta           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenJet_mass          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenJet_phi           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenJet_pt            | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "nGenPart             | uint32_t                 | AsDtype('>u4')\n",
      "GenPart_eta          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenPart_mass         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenPart_phi          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenPart_pt           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenPart_genPartId... | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "GenPart_pdgId        | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "GenPart_status       | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "GenPart_statusFlags  | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "nSubGenJetAK8        | uint32_t                 | AsDtype('>u4')\n",
      "SubGenJetAK8_eta     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SubGenJetAK8_mass    | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SubGenJetAK8_phi     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SubGenJetAK8_pt      | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Generator_binvar     | float                    | AsDtype('>f4')\n",
      "Generator_scalePDF   | float                    | AsDtype('>f4')\n",
      "Generator_weight     | float                    | AsDtype('>f4')\n",
      "Generator_x1         | float                    | AsDtype('>f4')\n",
      "Generator_x2         | float                    | AsDtype('>f4')\n",
      "Generator_xpdf1      | float                    | AsDtype('>f4')\n",
      "Generator_xpdf2      | float                    | AsDtype('>f4')\n",
      "Generator_id1        | int32_t                  | AsDtype('>i4')\n",
      "Generator_id2        | int32_t                  | AsDtype('>i4')\n",
      "GenVtx_x             | float                    | AsDtype('>f4')\n",
      "GenVtx_y             | float                    | AsDtype('>f4')\n",
      "GenVtx_z             | float                    | AsDtype('>f4')\n",
      "nGenVisTau           | uint32_t                 | AsDtype('>u4')\n",
      "GenVisTau_eta        | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenVisTau_mass       | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenVisTau_phi        | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenVisTau_pt         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenVisTau_charge     | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "GenVisTau_genPart... | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "GenVisTau_status     | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "genWeight            | float                    | AsDtype('>f4')\n",
      "LHEWeight_origina... | float                    | AsDtype('>f4')\n",
      "nLHEPdfWeight        | uint32_t                 | AsDtype('>u4')\n",
      "LHEPdfWeight         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "nLHEReweightingWe... | uint32_t                 | AsDtype('>u4')\n",
      "LHEReweightingWeight | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "nLHEScaleWeight      | uint32_t                 | AsDtype('>u4')\n",
      "LHEScaleWeight       | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "nPSWeight            | uint32_t                 | AsDtype('>u4')\n",
      "PSWeight             | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "nIsoTrack            | uint32_t                 | AsDtype('>u4')\n",
      "IsoTrack_dxy         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "IsoTrack_dz          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "IsoTrack_eta         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "IsoTrack_pfRelIso... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "IsoTrack_pfRelIso... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "IsoTrack_phi         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "IsoTrack_pt          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "IsoTrack_miniPFRe... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "IsoTrack_miniPFRe... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "IsoTrack_charge      | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "IsoTrack_fromPV      | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "IsoTrack_pdgId       | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "IsoTrack_isHighPu... | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "IsoTrack_isPFcand    | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "IsoTrack_isFromLo... | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "nJet                 | uint32_t                 | AsDtype('>u4')\n",
      "Jet_area             | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_btagCSVV2        | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_btagDeepB        | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_btagDeepCvB      | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_btagDeepCvL      | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_btagDeepFlavB    | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_btagDeepFlavCvB  | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_btagDeepFlavCvL  | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_btagDeepFlavQG   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_chEmEF           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_chFPV0EF         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_chHEF            | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_eta              | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_hfsigmaEtaEta    | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_hfsigmaPhiPhi    | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_mass             | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_muEF             | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_muonSubtrFactor  | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_neEmEF           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_neHEF            | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_phi              | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_pt               | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_puIdDisc         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_qgl              | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_rawFactor        | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_bRegCorr         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_bRegRes          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_cRegCorr         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_cRegRes          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Jet_electronIdx1     | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Jet_electronIdx2     | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Jet_hfadjacentEta... | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Jet_hfcentralEtaS... | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Jet_jetId            | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Jet_muonIdx1         | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Jet_muonIdx2         | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Jet_nElectrons       | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Jet_nMuons           | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Jet_puId             | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Jet_nConstituents    | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "L1PreFiringWeight_Dn | float                    | AsDtype('>f4')\n",
      "L1PreFiringWeight... | float                    | AsDtype('>f4')\n",
      "L1PreFiringWeight... | float                    | AsDtype('>f4')\n",
      "L1PreFiringWeight... | float                    | AsDtype('>f4')\n",
      "L1PreFiringWeight... | float                    | AsDtype('>f4')\n",
      "L1PreFiringWeight... | float                    | AsDtype('>f4')\n",
      "L1PreFiringWeight... | float                    | AsDtype('>f4')\n",
      "L1PreFiringWeight... | float                    | AsDtype('>f4')\n",
      "L1PreFiringWeight... | float                    | AsDtype('>f4')\n",
      "L1PreFiringWeight... | float                    | AsDtype('>f4')\n",
      "L1PreFiringWeight_Up | float                    | AsDtype('>f4')\n",
      "LHE_HT               | float                    | AsDtype('>f4')\n",
      "LHE_HTIncoming       | float                    | AsDtype('>f4')\n",
      "LHE_Vpt              | float                    | AsDtype('>f4')\n",
      "LHE_AlphaS           | float                    | AsDtype('>f4')\n",
      "LHE_Njets            | uint8_t                  | AsDtype('uint8')\n",
      "LHE_Nb               | uint8_t                  | AsDtype('uint8')\n",
      "LHE_Nc               | uint8_t                  | AsDtype('uint8')\n",
      "LHE_Nuds             | uint8_t                  | AsDtype('uint8')\n",
      "LHE_Nglu             | uint8_t                  | AsDtype('uint8')\n",
      "LHE_NpNLO            | uint8_t                  | AsDtype('uint8')\n",
      "LHE_NpLO             | uint8_t                  | AsDtype('uint8')\n",
      "nLHEPart             | uint32_t                 | AsDtype('>u4')\n",
      "LHEPart_pt           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LHEPart_eta          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LHEPart_phi          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LHEPart_mass         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LHEPart_incomingpz   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LHEPart_pdgId        | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "LHEPart_status       | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "LHEPart_spin         | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "nLowPtElectron       | uint32_t                 | AsDtype('>u4')\n",
      "LowPtElectron_ID     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_con... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_del... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_dxy    | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_dxyErr | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_dz     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_dzErr  | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_eIn... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_emb... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_ene... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_eta    | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_hoe    | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_mass   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_min... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_min... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_phi    | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_pt     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_ptb... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_r9     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_scE... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_sieie  | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_unb... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "LowPtElectron_charge | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "LowPtElectron_convWP | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "LowPtElectron_pdgId  | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "LowPtElectron_con... | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "LowPtElectron_los... | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "GenMET_phi           | float                    | AsDtype('>f4')\n",
      "GenMET_pt            | float                    | AsDtype('>f4')\n",
      "MET_MetUnclustEnU... | float                    | AsDtype('>f4')\n",
      "MET_MetUnclustEnU... | float                    | AsDtype('>f4')\n",
      "MET_covXX            | float                    | AsDtype('>f4')\n",
      "MET_covXY            | float                    | AsDtype('>f4')\n",
      "MET_covYY            | float                    | AsDtype('>f4')\n",
      "MET_phi              | float                    | AsDtype('>f4')\n",
      "MET_pt               | float                    | AsDtype('>f4')\n",
      "MET_significance     | float                    | AsDtype('>f4')\n",
      "MET_sumEt            | float                    | AsDtype('>f4')\n",
      "MET_sumPtUnclustered | float                    | AsDtype('>f4')\n",
      "nMuon                | uint32_t                 | AsDtype('>u4')\n",
      "Muon_dxy             | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_dxyErr          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_dxybs           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_dz              | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_dzErr           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_eta             | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_ip3d            | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_jetPtRelv2      | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_jetRelIso       | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_mass            | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_miniPFRelIso... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_miniPFRelIso... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_pfRelIso03_all  | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_pfRelIso03_chg  | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_pfRelIso04_all  | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_phi             | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_pt              | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_ptErr           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_segmentComp     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_sip3d           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_softMva         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_tkRelIso        | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_tunepRelPt      | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_mvaLowPt        | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_mvaTTH          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Muon_charge          | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Muon_jetIdx          | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Muon_nStations       | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Muon_nTrackerLayers  | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Muon_pdgId           | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Muon_tightCharge     | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Muon_fsrPhotonIdx    | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Muon_highPtId        | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Muon_highPurity      | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Muon_inTimeMuon      | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Muon_isGlobal        | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Muon_isPFcand        | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Muon_isStandalone    | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Muon_isTracker       | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Muon_jetNDauCharged  | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Muon_looseId         | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Muon_mediumId        | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Muon_mediumPromptId  | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Muon_miniIsoId       | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Muon_multiIsoId      | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Muon_mvaId           | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Muon_mvaLowPtId      | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Muon_pfIsoId         | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Muon_puppiIsoId      | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Muon_softId          | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Muon_softMvaId       | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Muon_tightId         | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Muon_tkIsoId         | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Muon_triggerIdLoose  | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "nPhoton              | uint32_t                 | AsDtype('>u4')\n",
      "Photon_dEscaleDown   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Photon_dEscaleUp     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Photon_dEsigmaDown   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Photon_dEsigmaUp     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Photon_eCorr         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Photon_energyErr     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Photon_eta           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Photon_hoe           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Photon_mass          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Photon_mvaID         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Photon_mvaID_Fall... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Photon_pfRelIso03... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Photon_pfRelIso03... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Photon_phi           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Photon_pt            | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Photon_r9            | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Photon_sieie         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Photon_charge        | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Photon_cutBased      | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Photon_cutBased_F... | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Photon_electronIdx   | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Photon_jetIdx        | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Photon_pdgId         | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Photon_vidNestedW... | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Photon_electronVeto  | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Photon_isScEtaEB     | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Photon_isScEtaEE     | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Photon_mvaID_WP80    | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Photon_mvaID_WP90    | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Photon_pixelSeed     | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Photon_seedGain      | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Pileup_nTrueInt      | float                    | AsDtype('>f4')\n",
      "Pileup_pudensity     | float                    | AsDtype('>f4')\n",
      "Pileup_gpudensity    | float                    | AsDtype('>f4')\n",
      "Pileup_nPU           | int32_t                  | AsDtype('>i4')\n",
      "Pileup_sumEOOT       | int32_t                  | AsDtype('>i4')\n",
      "Pileup_sumLOOT       | int32_t                  | AsDtype('>i4')\n",
      "PuppiMET_phi         | float                    | AsDtype('>f4')\n",
      "PuppiMET_phiJERDown  | float                    | AsDtype('>f4')\n",
      "PuppiMET_phiJERUp    | float                    | AsDtype('>f4')\n",
      "PuppiMET_phiJESDown  | float                    | AsDtype('>f4')\n",
      "PuppiMET_phiJESUp    | float                    | AsDtype('>f4')\n",
      "PuppiMET_phiUnclu... | float                    | AsDtype('>f4')\n",
      "PuppiMET_phiUnclu... | float                    | AsDtype('>f4')\n",
      "PuppiMET_pt          | float                    | AsDtype('>f4')\n",
      "PuppiMET_ptJERDown   | float                    | AsDtype('>f4')\n",
      "PuppiMET_ptJERUp     | float                    | AsDtype('>f4')\n",
      "PuppiMET_ptJESDown   | float                    | AsDtype('>f4')\n",
      "PuppiMET_ptJESUp     | float                    | AsDtype('>f4')\n",
      "PuppiMET_ptUnclus... | float                    | AsDtype('>f4')\n",
      "PuppiMET_ptUnclus... | float                    | AsDtype('>f4')\n",
      "PuppiMET_sumEt       | float                    | AsDtype('>f4')\n",
      "RawMET_phi           | float                    | AsDtype('>f4')\n",
      "RawMET_pt            | float                    | AsDtype('>f4')\n",
      "RawMET_sumEt         | float                    | AsDtype('>f4')\n",
      "RawPuppiMET_phi      | float                    | AsDtype('>f4')\n",
      "RawPuppiMET_pt       | float                    | AsDtype('>f4')\n",
      "RawPuppiMET_sumEt    | float                    | AsDtype('>f4')\n",
      "fixedGridRhoFastj... | float                    | AsDtype('>f4')\n",
      "fixedGridRhoFastj... | float                    | AsDtype('>f4')\n",
      "fixedGridRhoFastj... | float                    | AsDtype('>f4')\n",
      "fixedGridRhoFastj... | float                    | AsDtype('>f4')\n",
      "fixedGridRhoFastj... | float                    | AsDtype('>f4')\n",
      "nGenDressedLepton    | uint32_t                 | AsDtype('>u4')\n",
      "GenDressedLepton_eta | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenDressedLepton_... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenDressedLepton_phi | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenDressedLepton_pt  | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenDressedLepton_... | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "GenDressedLepton_... | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "nGenIsolatedPhoton   | uint32_t                 | AsDtype('>u4')\n",
      "GenIsolatedPhoton... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenIsolatedPhoton... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenIsolatedPhoton... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "GenIsolatedPhoton_pt | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "nSoftActivityJet     | uint32_t                 | AsDtype('>u4')\n",
      "SoftActivityJet_eta  | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SoftActivityJet_phi  | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SoftActivityJet_pt   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SoftActivityJetHT    | float                    | AsDtype('>f4')\n",
      "SoftActivityJetHT10  | float                    | AsDtype('>f4')\n",
      "SoftActivityJetHT2   | float                    | AsDtype('>f4')\n",
      "SoftActivityJetHT5   | float                    | AsDtype('>f4')\n",
      "SoftActivityJetNj... | int32_t                  | AsDtype('>i4')\n",
      "SoftActivityJetNj... | int32_t                  | AsDtype('>i4')\n",
      "SoftActivityJetNj... | int32_t                  | AsDtype('>i4')\n",
      "nSubJet              | uint32_t                 | AsDtype('>u4')\n",
      "SubJet_btagCSVV2     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SubJet_btagDeepB     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SubJet_eta           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SubJet_mass          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SubJet_n2b1          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SubJet_n3b1          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SubJet_phi           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SubJet_pt            | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SubJet_rawFactor     | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SubJet_tau1          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SubJet_tau2          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SubJet_tau3          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SubJet_tau4          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "nTau                 | uint32_t                 | AsDtype('>u4')\n",
      "Tau_chargedIso       | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_dxy              | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_dz               | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_eta              | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_leadTkDeltaEta   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_leadTkDeltaPhi   | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_leadTkPtOverT... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_mass             | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_neutralIso       | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_phi              | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_photonsOutsid... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_pt               | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_puCorr           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_rawDeepTau201... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_rawDeepTau201... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_rawDeepTau201... | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_rawIso           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_rawIsodR03       | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "Tau_charge           | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Tau_decayMode        | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Tau_jetIdx           | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Tau_idAntiEleDead... | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Tau_idAntiMu         | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Tau_idDecayModeOl... | bool[]                   | AsJagged(AsDtype('bool'))\n",
      "Tau_idDeepTau2017... | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Tau_idDeepTau2017... | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Tau_idDeepTau2017... | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "TkMET_phi            | float                    | AsDtype('>f4')\n",
      "TkMET_pt             | float                    | AsDtype('>f4')\n",
      "TkMET_sumEt          | float                    | AsDtype('>f4')\n",
      "nTrigObj             | uint32_t                 | AsDtype('>u4')\n",
      "TrigObj_pt           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "TrigObj_eta          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "TrigObj_phi          | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "TrigObj_l1pt         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "TrigObj_l1pt_2       | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "TrigObj_l2pt         | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "TrigObj_id           | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "TrigObj_l1iso        | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "TrigObj_l1charge     | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "TrigObj_filterBits   | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "genTtbarId           | int32_t                  | AsDtype('>i4')\n",
      "nOtherPV             | uint32_t                 | AsDtype('>u4')\n",
      "OtherPV_z            | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "PV_ndof              | float                    | AsDtype('>f4')\n",
      "PV_x                 | float                    | AsDtype('>f4')\n",
      "PV_y                 | float                    | AsDtype('>f4')\n",
      "PV_z                 | float                    | AsDtype('>f4')\n",
      "PV_chi2              | float                    | AsDtype('>f4')\n",
      "PV_score             | float                    | AsDtype('>f4')\n",
      "PV_npvs              | int32_t                  | AsDtype('>i4')\n",
      "PV_npvsGood          | int32_t                  | AsDtype('>i4')\n",
      "nSV                  | uint32_t                 | AsDtype('>u4')\n",
      "SV_dlen              | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SV_dlenSig           | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SV_dxy               | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SV_dxySig            | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SV_pAngle            | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SV_charge            | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "boostedTau_genPar... | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "boostedTau_genPar... | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Electron_genPartIdx  | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Electron_genPartFlav | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "FatJet_genJetAK8Idx  | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "FatJet_hadronFlavour | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "FatJet_nBHadrons     | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "FatJet_nCHadrons     | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "GenJetAK8_partonF... | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "GenJetAK8_hadronF... | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "GenJet_partonFlavour | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "GenJet_hadronFlavour | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "GenVtx_t0            | float                    | AsDtype('>f4')\n",
      "Jet_genJetIdx        | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Jet_hadronFlavour    | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Jet_partonFlavour    | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "LowPtElectron_gen... | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "LowPtElectron_gen... | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Muon_genPartIdx      | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Muon_genPartFlav     | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Photon_genPartIdx    | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Photon_genPartFlav   | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "MET_fiducialGenPhi   | float                    | AsDtype('>f4')\n",
      "MET_fiducialGenPt    | float                    | AsDtype('>f4')\n",
      "Electron_cleanmask   | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Jet_cleanmask        | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Muon_cleanmask       | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Photon_cleanmask     | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Tau_cleanmask        | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "SubJet_hadronFlavour | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "SubJet_nBHadrons     | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "SubJet_nCHadrons     | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "SV_chi2              | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SV_eta               | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SV_mass              | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SV_ndof              | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SV_phi               | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SV_pt                | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SV_x                 | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SV_y                 | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SV_z                 | float[]                  | AsJagged(AsDtype('>f4'))\n",
      "SV_ntracks           | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "Tau_genPartIdx       | int32_t[]                | AsJagged(AsDtype('>i4'))\n",
      "Tau_genPartFlav      | uint8_t[]                | AsJagged(AsDtype('uint8'))\n",
      "L1_AlwaysTrue        | bool                     | AsDtype('bool')\n",
      "L1_BPTX_AND_Ref1_VME | bool                     | AsDtype('bool')\n",
      "L1_BPTX_AND_Ref3_VME | bool                     | AsDtype('bool')\n",
      "L1_BPTX_AND_Ref4_VME | bool                     | AsDtype('bool')\n",
      "L1_BPTX_BeamGas_B... | bool                     | AsDtype('bool')\n",
      "L1_BPTX_BeamGas_B... | bool                     | AsDtype('bool')\n",
      "L1_BPTX_BeamGas_R... | bool                     | AsDtype('bool')\n",
      "L1_BPTX_BeamGas_R... | bool                     | AsDtype('bool')\n",
      "L1_BPTX_NotOR_VME    | bool                     | AsDtype('bool')\n",
      "L1_BPTX_OR_Ref3_VME  | bool                     | AsDtype('bool')\n",
      "L1_BPTX_OR_Ref4_VME  | bool                     | AsDtype('bool')\n",
      "L1_BPTX_RefAND_VME   | bool                     | AsDtype('bool')\n",
      "L1_BptxMinus         | bool                     | AsDtype('bool')\n",
      "L1_BptxOR            | bool                     | AsDtype('bool')\n",
      "L1_BptxPlus          | bool                     | AsDtype('bool')\n",
      "L1_BptxXOR           | bool                     | AsDtype('bool')\n",
      "L1_CDC_SingleMu_3... | bool                     | AsDtype('bool')\n",
      "L1_DoubleEG8er2p5... | bool                     | AsDtype('bool')\n",
      "L1_DoubleEG8er2p5... | bool                     | AsDtype('bool')\n",
      "L1_DoubleEG8er2p5... | bool                     | AsDtype('bool')\n",
      "L1_DoubleEG8er2p5... | bool                     | AsDtype('bool')\n",
      "L1_DoubleEG8er2p5... | bool                     | AsDtype('bool')\n",
      "L1_DoubleEG_15_10... | bool                     | AsDtype('bool')\n",
      "L1_DoubleEG_20_10... | bool                     | AsDtype('bool')\n",
      "L1_DoubleEG_22_10... | bool                     | AsDtype('bool')\n",
      "L1_DoubleEG_25_12... | bool                     | AsDtype('bool')\n",
      "L1_DoubleEG_25_14... | bool                     | AsDtype('bool')\n",
      "L1_DoubleEG_27_14... | bool                     | AsDtype('bool')\n",
      "L1_DoubleEG_Loose... | bool                     | AsDtype('bool')\n",
      "L1_DoubleEG_Loose... | bool                     | AsDtype('bool')\n",
      "L1_DoubleEG_Loose... | bool                     | AsDtype('bool')\n",
      "L1_DoubleEG_Loose... | bool                     | AsDtype('bool')\n",
      "L1_DoubleIsoTau32... | bool                     | AsDtype('bool')\n",
      "L1_DoubleIsoTau34... | bool                     | AsDtype('bool')\n",
      "L1_DoubleIsoTau36... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet100er... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet100er2p5 | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet112er... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet120er2p5 | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet150er2p5 | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet30er2... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet30er2... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet30er2... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet30er2... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet30er2... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet30er2... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet35_Ma... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet40er2p5  | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet_100_... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet_110_... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet_115_... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet_115_... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet_120_... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet_120_... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet_80_3... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet_80_3... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet_80_3... | bool                     | AsDtype('bool')\n",
      "L1_DoubleJet_90_3... | bool                     | AsDtype('bool')\n",
      "L1_DoubleLooseIso... | bool                     | AsDtype('bool')\n",
      "L1_DoubleLooseIso... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu0         | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu0_Mass... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu0_OQ      | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu0_SQ      | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu0_SQ_OS   | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu0_dR_M... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu0er1p4... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu0er1p5_SQ | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu0er1p5... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu0er1p5... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu0er1p5... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu0er2p0... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu0er2p0... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu10_SQ     | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu18er2p1   | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu3_OS_D... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu3_SQ_E... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu3_SQ_E... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu3_SQ_E... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu3_SQ_E... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu3_SQ_H... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu3_SQ_H... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu3_SQ_H... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu3_dR_M... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu4_SQ_E... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu4_SQ_OS   | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu4_SQ_O... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu4p5_SQ_OS | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu4p5_SQ... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu4p5er2... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu4p5er2... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu5Upsil... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu5_SQ_E... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu9_SQ      | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu_12_5     | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu_15_5_SQ  | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu_15_7     | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu_15_7_... | bool                     | AsDtype('bool')\n",
      "L1_DoubleMu_15_7_SQ  | bool                     | AsDtype('bool')\n",
      "L1_DoubleTau70er2p1  | bool                     | AsDtype('bool')\n",
      "L1_ETM120            | bool                     | AsDtype('bool')\n",
      "L1_ETM150            | bool                     | AsDtype('bool')\n",
      "L1_ETMHF100          | bool                     | AsDtype('bool')\n",
      "L1_ETMHF100_HTT60er  | bool                     | AsDtype('bool')\n",
      "L1_ETMHF110          | bool                     | AsDtype('bool')\n",
      "L1_ETMHF110_HTT60er  | bool                     | AsDtype('bool')\n",
      "L1_ETMHF110_HTT60... | bool                     | AsDtype('bool')\n",
      "L1_ETMHF120          | bool                     | AsDtype('bool')\n",
      "L1_ETMHF120_HTT60er  | bool                     | AsDtype('bool')\n",
      "L1_ETMHF120_NotSe... | bool                     | AsDtype('bool')\n",
      "L1_ETMHF130          | bool                     | AsDtype('bool')\n",
      "L1_ETMHF130_HTT60er  | bool                     | AsDtype('bool')\n",
      "L1_ETMHF140          | bool                     | AsDtype('bool')\n",
      "L1_ETMHF150          | bool                     | AsDtype('bool')\n",
      "L1_ETMHF90_HTT60er   | bool                     | AsDtype('bool')\n",
      "L1_ETT1200           | bool                     | AsDtype('bool')\n",
      "L1_ETT1600           | bool                     | AsDtype('bool')\n",
      "L1_ETT2000           | bool                     | AsDtype('bool')\n",
      "L1_FirstBunchAfte... | bool                     | AsDtype('bool')\n",
      "L1_FirstBunchBefo... | bool                     | AsDtype('bool')\n",
      "L1_FirstBunchInTrain | bool                     | AsDtype('bool')\n",
      "L1_FirstCollision... | bool                     | AsDtype('bool')\n",
      "L1_FirstCollision... | bool                     | AsDtype('bool')\n",
      "L1_HCAL_LaserMon_... | bool                     | AsDtype('bool')\n",
      "L1_HCAL_LaserMon_... | bool                     | AsDtype('bool')\n",
      "L1_HTT120er          | bool                     | AsDtype('bool')\n",
      "L1_HTT160er          | bool                     | AsDtype('bool')\n",
      "L1_HTT200er          | bool                     | AsDtype('bool')\n",
      "L1_HTT255er          | bool                     | AsDtype('bool')\n",
      "L1_HTT280er          | bool                     | AsDtype('bool')\n",
      "L1_HTT280er_QuadJ... | bool                     | AsDtype('bool')\n",
      "L1_HTT320er          | bool                     | AsDtype('bool')\n",
      "L1_HTT320er_QuadJ... | bool                     | AsDtype('bool')\n",
      "L1_HTT320er_QuadJ... | bool                     | AsDtype('bool')\n",
      "L1_HTT320er_QuadJ... | bool                     | AsDtype('bool')\n",
      "L1_HTT360er          | bool                     | AsDtype('bool')\n",
      "L1_HTT400er          | bool                     | AsDtype('bool')\n",
      "L1_HTT450er          | bool                     | AsDtype('bool')\n",
      "L1_IsoEG32er2p5_Mt40 | bool                     | AsDtype('bool')\n",
      "L1_IsoEG32er2p5_Mt44 | bool                     | AsDtype('bool')\n",
      "L1_IsoEG32er2p5_Mt48 | bool                     | AsDtype('bool')\n",
      "L1_IsoTau40er2p1_... | bool                     | AsDtype('bool')\n",
      "L1_IsoTau40er2p1_... | bool                     | AsDtype('bool')\n",
      "L1_IsoTau40er2p1_... | bool                     | AsDtype('bool')\n",
      "L1_IsoTau40er2p1_... | bool                     | AsDtype('bool')\n",
      "L1_IsolatedBunch     | bool                     | AsDtype('bool')\n",
      "L1_LastBunchInTrain  | bool                     | AsDtype('bool')\n",
      "L1_LastCollisionI... | bool                     | AsDtype('bool')\n",
      "L1_LooseIsoEG22er... | bool                     | AsDtype('bool')\n",
      "L1_LooseIsoEG22er... | bool                     | AsDtype('bool')\n",
      "L1_LooseIsoEG24er... | bool                     | AsDtype('bool')\n",
      "L1_LooseIsoEG24er... | bool                     | AsDtype('bool')\n",
      "L1_LooseIsoEG26er... | bool                     | AsDtype('bool')\n",
      "L1_LooseIsoEG26er... | bool                     | AsDtype('bool')\n",
      "L1_LooseIsoEG28er... | bool                     | AsDtype('bool')\n",
      "L1_LooseIsoEG28er... | bool                     | AsDtype('bool')\n",
      "L1_LooseIsoEG30er... | bool                     | AsDtype('bool')\n",
      "L1_LooseIsoEG30er... | bool                     | AsDtype('bool')\n",
      "L1_MinimumBiasHF0... | bool                     | AsDtype('bool')\n",
      "L1_Mu10er2p3_Jet3... | bool                     | AsDtype('bool')\n",
      "L1_Mu12er2p3_Jet4... | bool                     | AsDtype('bool')\n",
      "L1_Mu12er2p3_Jet4... | bool                     | AsDtype('bool')\n",
      "L1_Mu18er2p1_Tau2... | bool                     | AsDtype('bool')\n",
      "L1_Mu18er2p1_Tau2... | bool                     | AsDtype('bool')\n",
      "L1_Mu20_EG10er2p5    | bool                     | AsDtype('bool')\n",
      "L1_Mu22er2p1_IsoT... | bool                     | AsDtype('bool')\n",
      "L1_Mu22er2p1_IsoT... | bool                     | AsDtype('bool')\n",
      "L1_Mu22er2p1_IsoT... | bool                     | AsDtype('bool')\n",
      "L1_Mu22er2p1_IsoT... | bool                     | AsDtype('bool')\n",
      "L1_Mu22er2p1_Tau7... | bool                     | AsDtype('bool')\n",
      "L1_Mu3_Jet120er2p... | bool                     | AsDtype('bool')\n",
      "L1_Mu3_Jet120er2p... | bool                     | AsDtype('bool')\n",
      "L1_Mu3_Jet16er2p5... | bool                     | AsDtype('bool')\n",
      "L1_Mu3_Jet30er2p5    | bool                     | AsDtype('bool')\n",
      "L1_Mu3_Jet35er2p5... | bool                     | AsDtype('bool')\n",
      "L1_Mu3_Jet60er2p5... | bool                     | AsDtype('bool')\n",
      "L1_Mu3_Jet80er2p5... | bool                     | AsDtype('bool')\n",
      "L1_Mu3er1p5_Jet10... | bool                     | AsDtype('bool')\n",
      "L1_Mu3er1p5_Jet10... | bool                     | AsDtype('bool')\n",
      "L1_Mu5_EG23er2p5     | bool                     | AsDtype('bool')\n",
      "L1_Mu5_LooseIsoEG... | bool                     | AsDtype('bool')\n",
      "L1_Mu6_DoubleEG10... | bool                     | AsDtype('bool')\n",
      "L1_Mu6_DoubleEG12... | bool                     | AsDtype('bool')\n",
      "L1_Mu6_DoubleEG15... | bool                     | AsDtype('bool')\n",
      "L1_Mu6_DoubleEG17... | bool                     | AsDtype('bool')\n",
      "L1_Mu6_HTT240er      | bool                     | AsDtype('bool')\n",
      "L1_Mu6_HTT250er      | bool                     | AsDtype('bool')\n",
      "L1_Mu7_EG23er2p5     | bool                     | AsDtype('bool')\n",
      "L1_Mu7_LooseIsoEG... | bool                     | AsDtype('bool')\n",
      "L1_Mu7_LooseIsoEG... | bool                     | AsDtype('bool')\n",
      "L1_NotBptxOR         | bool                     | AsDtype('bool')\n",
      "L1_QuadJet36er2p5... | bool                     | AsDtype('bool')\n",
      "L1_QuadJet60er2p5    | bool                     | AsDtype('bool')\n",
      "L1_QuadJet_95_75_... | bool                     | AsDtype('bool')\n",
      "L1_QuadMu0           | bool                     | AsDtype('bool')\n",
      "L1_QuadMu0_OQ        | bool                     | AsDtype('bool')\n",
      "L1_QuadMu0_SQ        | bool                     | AsDtype('bool')\n",
      "L1_SecondBunchInT... | bool                     | AsDtype('bool')\n",
      "L1_SecondLastBunc... | bool                     | AsDtype('bool')\n",
      "L1_SingleEG10er2p5   | bool                     | AsDtype('bool')\n",
      "L1_SingleEG15er2p5   | bool                     | AsDtype('bool')\n",
      "L1_SingleEG26er2p5   | bool                     | AsDtype('bool')\n",
      "L1_SingleEG34er2p5   | bool                     | AsDtype('bool')\n",
      "L1_SingleEG36er2p5   | bool                     | AsDtype('bool')\n",
      "L1_SingleEG38er2p5   | bool                     | AsDtype('bool')\n",
      "L1_SingleEG40er2p5   | bool                     | AsDtype('bool')\n",
      "L1_SingleEG42er2p5   | bool                     | AsDtype('bool')\n",
      "L1_SingleEG45er2p5   | bool                     | AsDtype('bool')\n",
      "L1_SingleEG50        | bool                     | AsDtype('bool')\n",
      "L1_SingleEG60        | bool                     | AsDtype('bool')\n",
      "L1_SingleEG8er2p5    | bool                     | AsDtype('bool')\n",
      "L1_SingleIsoEG24e... | bool                     | AsDtype('bool')\n",
      "L1_SingleIsoEG24e... | bool                     | AsDtype('bool')\n",
      "L1_SingleIsoEG26e... | bool                     | AsDtype('bool')\n",
      "L1_SingleIsoEG26e... | bool                     | AsDtype('bool')\n",
      "L1_SingleIsoEG26e... | bool                     | AsDtype('bool')\n",
      "L1_SingleIsoEG28e... | bool                     | AsDtype('bool')\n",
      "L1_SingleIsoEG28e... | bool                     | AsDtype('bool')\n",
      "L1_SingleIsoEG28e... | bool                     | AsDtype('bool')\n",
      "L1_SingleIsoEG30e... | bool                     | AsDtype('bool')\n",
      "L1_SingleIsoEG30e... | bool                     | AsDtype('bool')\n",
      "L1_SingleIsoEG32e... | bool                     | AsDtype('bool')\n",
      "L1_SingleIsoEG32e... | bool                     | AsDtype('bool')\n",
      "L1_SingleIsoEG34e... | bool                     | AsDtype('bool')\n",
      "L1_SingleJet10erHE   | bool                     | AsDtype('bool')\n",
      "L1_SingleJet120      | bool                     | AsDtype('bool')\n",
      "L1_SingleJet120_F... | bool                     | AsDtype('bool')\n",
      "L1_SingleJet120er2p5 | bool                     | AsDtype('bool')\n",
      "L1_SingleJet12erHE   | bool                     | AsDtype('bool')\n",
      "L1_SingleJet140er2p5 | bool                     | AsDtype('bool')\n",
      "L1_SingleJet140er... | bool                     | AsDtype('bool')\n",
      "L1_SingleJet140er... | bool                     | AsDtype('bool')\n",
      "L1_SingleJet160er2p5 | bool                     | AsDtype('bool')\n",
      "L1_SingleJet180      | bool                     | AsDtype('bool')\n",
      "L1_SingleJet180er2p5 | bool                     | AsDtype('bool')\n",
      "L1_SingleJet200      | bool                     | AsDtype('bool')\n",
      "L1_SingleJet20er2... | bool                     | AsDtype('bool')\n",
      "L1_SingleJet20er2... | bool                     | AsDtype('bool')\n",
      "L1_SingleJet35       | bool                     | AsDtype('bool')\n",
      "L1_SingleJet35_FW... | bool                     | AsDtype('bool')\n",
      "L1_SingleJet35er2p5  | bool                     | AsDtype('bool')\n",
      "L1_SingleJet43er2... | bool                     | AsDtype('bool')\n",
      "L1_SingleJet46er2... | bool                     | AsDtype('bool')\n",
      "L1_SingleJet60       | bool                     | AsDtype('bool')\n",
      "L1_SingleJet60_FW... | bool                     | AsDtype('bool')\n",
      "L1_SingleJet60er2p5  | bool                     | AsDtype('bool')\n",
      "L1_SingleJet8erHE    | bool                     | AsDtype('bool')\n",
      "L1_SingleJet90       | bool                     | AsDtype('bool')\n",
      "L1_SingleJet90_FW... | bool                     | AsDtype('bool')\n",
      "L1_SingleJet90er2p5  | bool                     | AsDtype('bool')\n",
      "L1_SingleLooseIso... | bool                     | AsDtype('bool')\n",
      "L1_SingleLooseIso... | bool                     | AsDtype('bool')\n",
      "L1_SingleMu0_BMTF    | bool                     | AsDtype('bool')\n",
      "L1_SingleMu0_DQ      | bool                     | AsDtype('bool')\n",
      "L1_SingleMu0_EMTF    | bool                     | AsDtype('bool')\n",
      "L1_SingleMu0_OMTF    | bool                     | AsDtype('bool')\n",
      "L1_SingleMu10er1p5   | bool                     | AsDtype('bool')\n",
      "L1_SingleMu12_DQ_... | bool                     | AsDtype('bool')\n",
      "L1_SingleMu12_DQ_... | bool                     | AsDtype('bool')\n",
      "L1_SingleMu12_DQ_... | bool                     | AsDtype('bool')\n",
      "L1_SingleMu12er1p5   | bool                     | AsDtype('bool')\n",
      "L1_SingleMu14er1p5   | bool                     | AsDtype('bool')\n",
      "L1_SingleMu15_DQ     | bool                     | AsDtype('bool')\n",
      "L1_SingleMu16er1p5   | bool                     | AsDtype('bool')\n",
      "L1_SingleMu18        | bool                     | AsDtype('bool')\n",
      "L1_SingleMu18er1p5   | bool                     | AsDtype('bool')\n",
      "L1_SingleMu20        | bool                     | AsDtype('bool')\n",
      "L1_SingleMu22        | bool                     | AsDtype('bool')\n",
      "L1_SingleMu22_BMTF   | bool                     | AsDtype('bool')\n",
      "L1_SingleMu22_EMTF   | bool                     | AsDtype('bool')\n",
      "L1_SingleMu22_OMTF   | bool                     | AsDtype('bool')\n",
      "L1_SingleMu25        | bool                     | AsDtype('bool')\n",
      "L1_SingleMu3         | bool                     | AsDtype('bool')\n",
      "L1_SingleMu5         | bool                     | AsDtype('bool')\n",
      "L1_SingleMu6er1p5    | bool                     | AsDtype('bool')\n",
      "L1_SingleMu7         | bool                     | AsDtype('bool')\n",
      "L1_SingleMu7_DQ      | bool                     | AsDtype('bool')\n",
      "L1_SingleMu7er1p5    | bool                     | AsDtype('bool')\n",
      "L1_SingleMu8er1p5    | bool                     | AsDtype('bool')\n",
      "L1_SingleMu9er1p5    | bool                     | AsDtype('bool')\n",
      "L1_SingleMuCosmics   | bool                     | AsDtype('bool')\n",
      "L1_SingleMuCosmic... | bool                     | AsDtype('bool')\n",
      "L1_SingleMuCosmic... | bool                     | AsDtype('bool')\n",
      "L1_SingleMuCosmic... | bool                     | AsDtype('bool')\n",
      "L1_SingleMuOpen      | bool                     | AsDtype('bool')\n",
      "L1_SingleMuOpen_N... | bool                     | AsDtype('bool')\n",
      "L1_SingleMuOpen_e... | bool                     | AsDtype('bool')\n",
      "L1_SingleMuOpen_e... | bool                     | AsDtype('bool')\n",
      "L1_SingleTau120er2p1 | bool                     | AsDtype('bool')\n",
      "L1_SingleTau130er2p1 | bool                     | AsDtype('bool')\n",
      "L1_TOTEM_1           | bool                     | AsDtype('bool')\n",
      "L1_TOTEM_2           | bool                     | AsDtype('bool')\n",
      "L1_TOTEM_3           | bool                     | AsDtype('bool')\n",
      "L1_TOTEM_4           | bool                     | AsDtype('bool')\n",
      "L1_TripleEG16er2p5   | bool                     | AsDtype('bool')\n",
      "L1_TripleEG_16_12... | bool                     | AsDtype('bool')\n",
      "L1_TripleEG_16_15... | bool                     | AsDtype('bool')\n",
      "L1_TripleEG_18_17... | bool                     | AsDtype('bool')\n",
      "L1_TripleEG_18_18... | bool                     | AsDtype('bool')\n",
      "L1_TripleJet_100_... | bool                     | AsDtype('bool')\n",
      "L1_TripleJet_105_... | bool                     | AsDtype('bool')\n",
      "L1_TripleJet_95_7... | bool                     | AsDtype('bool')\n",
      "L1_TripleMu0         | bool                     | AsDtype('bool')\n",
      "L1_TripleMu0_OQ      | bool                     | AsDtype('bool')\n",
      "L1_TripleMu0_SQ      | bool                     | AsDtype('bool')\n",
      "L1_TripleMu3         | bool                     | AsDtype('bool')\n",
      "L1_TripleMu3_SQ      | bool                     | AsDtype('bool')\n",
      "L1_TripleMu_5SQ_3... | bool                     | AsDtype('bool')\n",
      "L1_TripleMu_5SQ_3... | bool                     | AsDtype('bool')\n",
      "L1_TripleMu_5SQ_3... | bool                     | AsDtype('bool')\n",
      "L1_TripleMu_5_3_3    | bool                     | AsDtype('bool')\n",
      "L1_TripleMu_5_3_3_SQ | bool                     | AsDtype('bool')\n",
      "L1_TripleMu_5_3p5... | bool                     | AsDtype('bool')\n",
      "L1_TripleMu_5_3p5... | bool                     | AsDtype('bool')\n",
      "L1_TripleMu_5_3p5... | bool                     | AsDtype('bool')\n",
      "L1_TripleMu_5_4_2... | bool                     | AsDtype('bool')\n",
      "L1_TripleMu_5_5_3    | bool                     | AsDtype('bool')\n",
      "L1_UnpairedBunchB... | bool                     | AsDtype('bool')\n",
      "L1_UnpairedBunchB... | bool                     | AsDtype('bool')\n",
      "L1_ZeroBias          | bool                     | AsDtype('bool')\n",
      "L1_ZeroBias_copy     | bool                     | AsDtype('bool')\n",
      "L1_UnprefireableE... | bool                     | AsDtype('bool')\n",
      "Flag_HBHENoiseFilter | bool                     | AsDtype('bool')\n",
      "Flag_HBHENoiseIso... | bool                     | AsDtype('bool')\n",
      "Flag_CSCTightHalo... | bool                     | AsDtype('bool')\n",
      "Flag_CSCTightHalo... | bool                     | AsDtype('bool')\n",
      "Flag_CSCTightHalo... | bool                     | AsDtype('bool')\n",
      "Flag_globalTightH... | bool                     | AsDtype('bool')\n",
      "Flag_globalSuperT... | bool                     | AsDtype('bool')\n",
      "Flag_HcalStripHal... | bool                     | AsDtype('bool')\n",
      "Flag_hcalLaserEve... | bool                     | AsDtype('bool')\n",
      "Flag_EcalDeadCell... | bool                     | AsDtype('bool')\n",
      "Flag_EcalDeadCell... | bool                     | AsDtype('bool')\n",
      "Flag_ecalBadCalib... | bool                     | AsDtype('bool')\n",
      "Flag_goodVertices    | bool                     | AsDtype('bool')\n",
      "Flag_eeBadScFilter   | bool                     | AsDtype('bool')\n",
      "Flag_ecalLaserCor... | bool                     | AsDtype('bool')\n",
      "Flag_trkPOGFilters   | bool                     | AsDtype('bool')\n",
      "Flag_chargedHadro... | bool                     | AsDtype('bool')\n",
      "Flag_muonBadTrack... | bool                     | AsDtype('bool')\n",
      "Flag_BadChargedCa... | bool                     | AsDtype('bool')\n",
      "Flag_BadPFMuonFilter | bool                     | AsDtype('bool')\n",
      "Flag_BadPFMuonDzF... | bool                     | AsDtype('bool')\n",
      "Flag_hfNoisyHitsF... | bool                     | AsDtype('bool')\n",
      "Flag_BadChargedCa... | bool                     | AsDtype('bool')\n",
      "Flag_BadPFMuonSum... | bool                     | AsDtype('bool')\n",
      "Flag_trkPOG_manys... | bool                     | AsDtype('bool')\n",
      "Flag_trkPOG_tooma... | bool                     | AsDtype('bool')\n",
      "Flag_trkPOG_logEr... | bool                     | AsDtype('bool')\n",
      "Flag_METFilters      | bool                     | AsDtype('bool')\n",
      "L1Reco_step          | bool                     | AsDtype('bool')\n",
      "HLTriggerFirstPath   | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet360_T... | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet380_T... | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet400_T... | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet420_T... | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFHT750_Tr... | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFHT800_Tr... | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFHT850_Tr... | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFHT900_Tr... | bool                     | AsDtype('bool')\n",
      "HLT_CaloJet500_No... | bool                     | AsDtype('bool')\n",
      "HLT_CaloJet550_No... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu5_Ups... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu3_Dou... | bool                     | AsDtype('bool')\n",
      "HLT_Trimuon5_3p5_... | bool                     | AsDtype('bool')\n",
      "HLT_TrimuonOpen_5... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleEle25_C... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleEle27_C... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleEle33_C... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleEle24_e... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleEle8_Ca... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleEle8_Ca... | bool                     | AsDtype('bool')\n",
      "HLT_Ele27_Ele37_C... | bool                     | AsDtype('bool')\n",
      "HLT_Mu27_Ele37_Ca... | bool                     | AsDtype('bool')\n",
      "HLT_Mu37_Ele27_Ca... | bool                     | AsDtype('bool')\n",
      "HLT_Mu37_TkMu27      | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu4_3_Bs   | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu4_3_Jpsi | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu4_Jps... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu4_Low... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu3_Trk... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu3_TkM... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu4_Psi... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu4_Mas... | bool                     | AsDtype('bool')\n",
      "HLT_Mu3_PFJet40      | bool                     | AsDtype('bool')\n",
      "HLT_Mu7p5_L2Mu2_Jpsi | bool                     | AsDtype('bool')\n",
      "HLT_Mu7p5_L2Mu2_U... | bool                     | AsDtype('bool')\n",
      "HLT_Mu7p5_Track2_... | bool                     | AsDtype('bool')\n",
      "HLT_Mu7p5_Track3p... | bool                     | AsDtype('bool')\n",
      "HLT_Mu7p5_Track7_... | bool                     | AsDtype('bool')\n",
      "HLT_Mu7p5_Track2_... | bool                     | AsDtype('bool')\n",
      "HLT_Mu7p5_Track3p... | bool                     | AsDtype('bool')\n",
      "HLT_Mu7p5_Track7_... | bool                     | AsDtype('bool')\n",
      "HLT_Mu3_L1SingleM... | bool                     | AsDtype('bool')\n",
      "HLT_DoublePhoton3... | bool                     | AsDtype('bool')\n",
      "HLT_DoublePhoton70   | bool                     | AsDtype('bool')\n",
      "HLT_DoublePhoton85   | bool                     | AsDtype('bool')\n",
      "HLT_Ele20_WPTight... | bool                     | AsDtype('bool')\n",
      "HLT_Ele15_WPLoose... | bool                     | AsDtype('bool')\n",
      "HLT_Ele17_WPLoose... | bool                     | AsDtype('bool')\n",
      "HLT_Ele20_WPLoose... | bool                     | AsDtype('bool')\n",
      "HLT_Ele20_eta2p1_... | bool                     | AsDtype('bool')\n",
      "HLT_DiEle27_WPTig... | bool                     | AsDtype('bool')\n",
      "HLT_Ele27_WPTight... | bool                     | AsDtype('bool')\n",
      "HLT_Ele28_WPTight... | bool                     | AsDtype('bool')\n",
      "HLT_Ele30_WPTight... | bool                     | AsDtype('bool')\n",
      "HLT_Ele32_WPTight... | bool                     | AsDtype('bool')\n",
      "HLT_Ele35_WPTight... | bool                     | AsDtype('bool')\n",
      "HLT_Ele35_WPTight... | bool                     | AsDtype('bool')\n",
      "HLT_Ele38_WPTight... | bool                     | AsDtype('bool')\n",
      "HLT_Ele40_WPTight... | bool                     | AsDtype('bool')\n",
      "HLT_Ele32_WPTight... | bool                     | AsDtype('bool')\n",
      "HLT_Ele24_eta2p1_... | bool                     | AsDtype('bool')\n",
      "HLT_Ele24_eta2p1_... | bool                     | AsDtype('bool')\n",
      "HLT_Ele24_eta2p1_... | bool                     | AsDtype('bool')\n",
      "HLT_Ele24_eta2p1_... | bool                     | AsDtype('bool')\n",
      "HLT_Ele24_eta2p1_... | bool                     | AsDtype('bool')\n",
      "HLT_Ele24_eta2p1_... | bool                     | AsDtype('bool')\n",
      "HLT_HT450_Beamspot   | bool                     | AsDtype('bool')\n",
      "HLT_HT300_Beamspot   | bool                     | AsDtype('bool')\n",
      "HLT_ZeroBias_Beam... | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu20_eta2p... | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu20_eta2p... | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu20_eta2p... | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu20_eta2p... | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu20_eta2p... | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu20_eta2p... | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu24_eta2p... | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu24_eta2p... | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu24_eta2p... | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu24_eta2p... | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu27_Loose... | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu27_Mediu... | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu27_Tight... | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu20          | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu24          | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu24_eta2p1   | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu27          | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu30          | bool                     | AsDtype('bool')\n",
      "HLT_UncorrectedJe... | bool                     | AsDtype('bool')\n",
      "HLT_UncorrectedJe... | bool                     | AsDtype('bool')\n",
      "HLT_UncorrectedJe... | bool                     | AsDtype('bool')\n",
      "HLT_UncorrectedJe... | bool                     | AsDtype('bool')\n",
      "HLT_L1SingleMu18     | bool                     | AsDtype('bool')\n",
      "HLT_L1SingleMu25     | bool                     | AsDtype('bool')\n",
      "HLT_L2Mu10           | bool                     | AsDtype('bool')\n",
      "HLT_L2Mu10_NoVert... | bool                     | AsDtype('bool')\n",
      "HLT_L2Mu10_NoVert... | bool                     | AsDtype('bool')\n",
      "HLT_L2Mu45_NoVert... | bool                     | AsDtype('bool')\n",
      "HLT_L2Mu40_NoVert... | bool                     | AsDtype('bool')\n",
      "HLT_L2Mu50           | bool                     | AsDtype('bool')\n",
      "HLT_L2Mu23NoVtx_2Cha | bool                     | AsDtype('bool')\n",
      "HLT_L2Mu23NoVtx_2... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleL2Mu30N... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleL2Mu30N... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleL2Mu50     | bool                     | AsDtype('bool')\n",
      "HLT_DoubleL2Mu23N... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleL2Mu23N... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleL2Mu25N... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleL2Mu25N... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleL2Mu25N... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleL2Mu23N... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleL2Mu23N... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleL2Mu25N... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleL2Mu25N... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleL2Mu25N... | bool                     | AsDtype('bool')\n",
      "HLT_Mu17_TrkIsoVV... | bool                     | AsDtype('bool')\n",
      "HLT_Mu19_TrkIsoVV... | bool                     | AsDtype('bool')\n",
      "HLT_Mu17_TrkIsoVV... | bool                     | AsDtype('bool')\n",
      "HLT_Mu19_TrkIsoVV... | bool                     | AsDtype('bool')\n",
      "HLT_Mu17_TrkIsoVV... | bool                     | AsDtype('bool')\n",
      "HLT_Mu19_TrkIsoVV... | bool                     | AsDtype('bool')\n",
      "HLT_Mu17_TrkIsoVV... | bool                     | AsDtype('bool')\n",
      "HLT_Mu19_TrkIsoVV... | bool                     | AsDtype('bool')\n",
      "HLT_Mu25_TkMu0_Onia  | bool                     | AsDtype('bool')\n",
      "HLT_Mu30_TkMu0_Psi   | bool                     | AsDtype('bool')\n",
      "HLT_Mu30_TkMu0_Up... | bool                     | AsDtype('bool')\n",
      "HLT_Mu20_TkMu0_Phi   | bool                     | AsDtype('bool')\n",
      "HLT_Mu25_TkMu0_Phi   | bool                     | AsDtype('bool')\n",
      "HLT_Mu12             | bool                     | AsDtype('bool')\n",
      "HLT_Mu15             | bool                     | AsDtype('bool')\n",
      "HLT_Mu20             | bool                     | AsDtype('bool')\n",
      "HLT_Mu27             | bool                     | AsDtype('bool')\n",
      "HLT_Mu50             | bool                     | AsDtype('bool')\n",
      "HLT_Mu55             | bool                     | AsDtype('bool')\n",
      "HLT_OldMu100         | bool                     | AsDtype('bool')\n",
      "HLT_TkMu100          | bool                     | AsDtype('bool')\n",
      "HLT_DiPFJetAve40     | bool                     | AsDtype('bool')\n",
      "HLT_DiPFJetAve60     | bool                     | AsDtype('bool')\n",
      "HLT_DiPFJetAve80     | bool                     | AsDtype('bool')\n",
      "HLT_DiPFJetAve140    | bool                     | AsDtype('bool')\n",
      "HLT_DiPFJetAve200    | bool                     | AsDtype('bool')\n",
      "HLT_DiPFJetAve260    | bool                     | AsDtype('bool')\n",
      "HLT_DiPFJetAve320    | bool                     | AsDtype('bool')\n",
      "HLT_DiPFJetAve400    | bool                     | AsDtype('bool')\n",
      "HLT_DiPFJetAve500    | bool                     | AsDtype('bool')\n",
      "HLT_DiPFJetAve60_... | bool                     | AsDtype('bool')\n",
      "HLT_DiPFJetAve80_... | bool                     | AsDtype('bool')\n",
      "HLT_DiPFJetAve100... | bool                     | AsDtype('bool')\n",
      "HLT_DiPFJetAve160... | bool                     | AsDtype('bool')\n",
      "HLT_DiPFJetAve220... | bool                     | AsDtype('bool')\n",
      "HLT_DiPFJetAve300... | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet15       | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet25       | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet40       | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet60       | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet80       | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet140      | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet200      | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet260      | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet320      | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet400      | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet450      | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet500      | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet550      | bool                     | AsDtype('bool')\n",
      "HLT_PFJet15          | bool                     | AsDtype('bool')\n",
      "HLT_PFJet25          | bool                     | AsDtype('bool')\n",
      "HLT_PFJet40          | bool                     | AsDtype('bool')\n",
      "HLT_PFJet60          | bool                     | AsDtype('bool')\n",
      "HLT_PFJet80          | bool                     | AsDtype('bool')\n",
      "HLT_PFJet140         | bool                     | AsDtype('bool')\n",
      "HLT_PFJet200         | bool                     | AsDtype('bool')\n",
      "HLT_PFJet260         | bool                     | AsDtype('bool')\n",
      "HLT_PFJet320         | bool                     | AsDtype('bool')\n",
      "HLT_PFJet400         | bool                     | AsDtype('bool')\n",
      "HLT_PFJet450         | bool                     | AsDtype('bool')\n",
      "HLT_PFJet500         | bool                     | AsDtype('bool')\n",
      "HLT_PFJet550         | bool                     | AsDtype('bool')\n",
      "HLT_PFJetFwd15       | bool                     | AsDtype('bool')\n",
      "HLT_PFJetFwd25       | bool                     | AsDtype('bool')\n",
      "HLT_PFJetFwd40       | bool                     | AsDtype('bool')\n",
      "HLT_PFJetFwd60       | bool                     | AsDtype('bool')\n",
      "HLT_PFJetFwd80       | bool                     | AsDtype('bool')\n",
      "HLT_PFJetFwd140      | bool                     | AsDtype('bool')\n",
      "HLT_PFJetFwd200      | bool                     | AsDtype('bool')\n",
      "HLT_PFJetFwd260      | bool                     | AsDtype('bool')\n",
      "HLT_PFJetFwd320      | bool                     | AsDtype('bool')\n",
      "HLT_PFJetFwd400      | bool                     | AsDtype('bool')\n",
      "HLT_PFJetFwd450      | bool                     | AsDtype('bool')\n",
      "HLT_PFJetFwd500      | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJetFwd15    | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJetFwd25    | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJetFwd40    | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJetFwd60    | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJetFwd80    | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJetFwd140   | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJetFwd200   | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJetFwd260   | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJetFwd320   | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJetFwd400   | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJetFwd450   | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJetFwd500   | bool                     | AsDtype('bool')\n",
      "HLT_PFHT180          | bool                     | AsDtype('bool')\n",
      "HLT_PFHT250          | bool                     | AsDtype('bool')\n",
      "HLT_PFHT370          | bool                     | AsDtype('bool')\n",
      "HLT_PFHT430          | bool                     | AsDtype('bool')\n",
      "HLT_PFHT510          | bool                     | AsDtype('bool')\n",
      "HLT_PFHT590          | bool                     | AsDtype('bool')\n",
      "HLT_PFHT680          | bool                     | AsDtype('bool')\n",
      "HLT_PFHT780          | bool                     | AsDtype('bool')\n",
      "HLT_PFHT890          | bool                     | AsDtype('bool')\n",
      "HLT_PFHT1050         | bool                     | AsDtype('bool')\n",
      "HLT_PFHT500_PFMET... | bool                     | AsDtype('bool')\n",
      "HLT_PFHT500_PFMET... | bool                     | AsDtype('bool')\n",
      "HLT_PFHT700_PFMET... | bool                     | AsDtype('bool')\n",
      "HLT_PFHT700_PFMET... | bool                     | AsDtype('bool')\n",
      "HLT_PFHT800_PFMET... | bool                     | AsDtype('bool')\n",
      "HLT_PFHT800_PFMET... | bool                     | AsDtype('bool')\n",
      "HLT_PFMET110_PFMH... | bool                     | AsDtype('bool')\n",
      "HLT_PFMET120_PFMH... | bool                     | AsDtype('bool')\n",
      "HLT_PFMET130_PFMH... | bool                     | AsDtype('bool')\n",
      "HLT_PFMET140_PFMH... | bool                     | AsDtype('bool')\n",
      "HLT_PFMET100_PFMH... | bool                     | AsDtype('bool')\n",
      "HLT_PFMET110_PFMH... | bool                     | AsDtype('bool')\n",
      "HLT_PFMET120_PFMH... | bool                     | AsDtype('bool')\n",
      "HLT_PFMET130_PFMH... | bool                     | AsDtype('bool')\n",
      "HLT_PFMET140_PFMH... | bool                     | AsDtype('bool')\n",
      "HLT_PFMET120_PFMH... | bool                     | AsDtype('bool')\n",
      "HLT_PFMETNoMu120_... | bool                     | AsDtype('bool')\n",
      "HLT_PFMETTypeOne1... | bool                     | AsDtype('bool')\n",
      "HLT_PFMETTypeOne1... | bool                     | AsDtype('bool')\n",
      "HLT_PFMETTypeOne1... | bool                     | AsDtype('bool')\n",
      "HLT_PFMETTypeOne1... | bool                     | AsDtype('bool')\n",
      "HLT_PFMETTypeOne1... | bool                     | AsDtype('bool')\n",
      "HLT_PFMETNoMu110_... | bool                     | AsDtype('bool')\n",
      "HLT_PFMETNoMu120_... | bool                     | AsDtype('bool')\n",
      "HLT_PFMETNoMu130_... | bool                     | AsDtype('bool')\n",
      "HLT_PFMETNoMu140_... | bool                     | AsDtype('bool')\n",
      "HLT_MonoCentralPF... | bool                     | AsDtype('bool')\n",
      "HLT_MonoCentralPF... | bool                     | AsDtype('bool')\n",
      "HLT_MonoCentralPF... | bool                     | AsDtype('bool')\n",
      "HLT_MonoCentralPF... | bool                     | AsDtype('bool')\n",
      "HLT_L1ETMHadSeeds    | bool                     | AsDtype('bool')\n",
      "HLT_CaloMHT90        | bool                     | AsDtype('bool')\n",
      "HLT_CaloMET80_Not... | bool                     | AsDtype('bool')\n",
      "HLT_CaloMET90_Not... | bool                     | AsDtype('bool')\n",
      "HLT_CaloMET100_No... | bool                     | AsDtype('bool')\n",
      "HLT_CaloMET110_No... | bool                     | AsDtype('bool')\n",
      "HLT_CaloMET250_No... | bool                     | AsDtype('bool')\n",
      "HLT_CaloMET70_HBH... | bool                     | AsDtype('bool')\n",
      "HLT_CaloMET80_HBH... | bool                     | AsDtype('bool')\n",
      "HLT_CaloMET90_HBH... | bool                     | AsDtype('bool')\n",
      "HLT_CaloMET100_HB... | bool                     | AsDtype('bool')\n",
      "HLT_CaloMET250_HB... | bool                     | AsDtype('bool')\n",
      "HLT_CaloMET300_HB... | bool                     | AsDtype('bool')\n",
      "HLT_CaloMET350_HB... | bool                     | AsDtype('bool')\n",
      "HLT_PFMET200_NotC... | bool                     | AsDtype('bool')\n",
      "HLT_PFMET200_HBHE... | bool                     | AsDtype('bool')\n",
      "HLT_PFMET250_HBHE... | bool                     | AsDtype('bool')\n",
      "HLT_PFMET300_HBHE... | bool                     | AsDtype('bool')\n",
      "HLT_PFMET200_HBHE... | bool                     | AsDtype('bool')\n",
      "HLT_PFMETTypeOne2... | bool                     | AsDtype('bool')\n",
      "HLT_MET105_IsoTrk50  | bool                     | AsDtype('bool')\n",
      "HLT_MET120_IsoTrk50  | bool                     | AsDtype('bool')\n",
      "HLT_SingleJet30_M... | bool                     | AsDtype('bool')\n",
      "HLT_Mu12_DoublePF... | bool                     | AsDtype('bool')\n",
      "HLT_Mu12_DoublePF... | bool                     | AsDtype('bool')\n",
      "HLT_Mu12_DoublePF... | bool                     | AsDtype('bool')\n",
      "HLT_Mu12_DoublePF... | bool                     | AsDtype('bool')\n",
      "HLT_Mu12_DoublePF... | bool                     | AsDtype('bool')\n",
      "HLT_Mu12_DoublePF... | bool                     | AsDtype('bool')\n",
      "HLT_Mu12_DoublePF... | bool                     | AsDtype('bool')\n",
      "HLT_DoublePFJets4... | bool                     | AsDtype('bool')\n",
      "HLT_DoublePFJets1... | bool                     | AsDtype('bool')\n",
      "HLT_DoublePFJets2... | bool                     | AsDtype('bool')\n",
      "HLT_DoublePFJets3... | bool                     | AsDtype('bool')\n",
      "HLT_DoublePFJets1... | bool                     | AsDtype('bool')\n",
      "HLT_DoublePFJets1... | bool                     | AsDtype('bool')\n",
      "HLT_Photon300_NoHE   | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_TrkIsoVVL    | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_DiEle12_C... | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_DiEle12_C... | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_Ele8_Calo... | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_Ele8_Calo... | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_TrkIsoVVL... | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_TrkIsoVVL... | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_TrkIsoVVL... | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_TrkIsoVVL... | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_TrkIsoVVL... | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_TrkIsoVVL... | bool                     | AsDtype('bool')\n",
      "HLT_Mu17_TrkIsoVVL   | bool                     | AsDtype('bool')\n",
      "HLT_Mu19_TrkIsoVVL   | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK4DiJ... | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK4DiJ... | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK4DiJ... | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK4DiJ... | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK4DiJ... | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK4Jet... | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK8DiJ... | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK8Jet... | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK8Jet... | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK4DiJ... | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK4DiJ... | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK4DiJ... | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK4DiJ... | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK4DiJ... | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK4Jet... | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK8DiJ... | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK8Jet... | bool                     | AsDtype('bool')\n",
      "HLT_BTagMu_AK8Jet... | bool                     | AsDtype('bool')\n",
      "HLT_Ele15_Ele8_Ca... | bool                     | AsDtype('bool')\n",
      "HLT_Ele23_Ele12_C... | bool                     | AsDtype('bool')\n",
      "HLT_Ele23_Ele12_C... | bool                     | AsDtype('bool')\n",
      "HLT_Mu23_TrkIsoVV... | bool                     | AsDtype('bool')\n",
      "HLT_Mu23_TrkIsoVV... | bool                     | AsDtype('bool')\n",
      "HLT_Mu12_TrkIsoVV... | bool                     | AsDtype('bool')\n",
      "HLT_Mu12_TrkIsoVV... | bool                     | AsDtype('bool')\n",
      "HLT_Mu12_DoublePh... | bool                     | AsDtype('bool')\n",
      "HLT_TriplePhoton_... | bool                     | AsDtype('bool')\n",
      "HLT_TriplePhoton_... | bool                     | AsDtype('bool')\n",
      "HLT_TriplePhoton_... | bool                     | AsDtype('bool')\n",
      "HLT_TriplePhoton_... | bool                     | AsDtype('bool')\n",
      "HLT_TriplePhoton_... | bool                     | AsDtype('bool')\n",
      "HLT_Photon20         | bool                     | AsDtype('bool')\n",
      "HLT_Photon33         | bool                     | AsDtype('bool')\n",
      "HLT_Photon50         | bool                     | AsDtype('bool')\n",
      "HLT_Photon75         | bool                     | AsDtype('bool')\n",
      "HLT_Photon90         | bool                     | AsDtype('bool')\n",
      "HLT_Photon120        | bool                     | AsDtype('bool')\n",
      "HLT_Photon150        | bool                     | AsDtype('bool')\n",
      "HLT_Photon175        | bool                     | AsDtype('bool')\n",
      "HLT_Photon200        | bool                     | AsDtype('bool')\n",
      "HLT_Photon100EB_T... | bool                     | AsDtype('bool')\n",
      "HLT_Photon110EB_T... | bool                     | AsDtype('bool')\n",
      "HLT_Photon120EB_T... | bool                     | AsDtype('bool')\n",
      "HLT_Photon100EBHE10  | bool                     | AsDtype('bool')\n",
      "HLT_Photon100EEHE10  | bool                     | AsDtype('bool')\n",
      "HLT_Photon100EE_T... | bool                     | AsDtype('bool')\n",
      "HLT_Photon50_R9Id... | bool                     | AsDtype('bool')\n",
      "HLT_Photon75_R9Id... | bool                     | AsDtype('bool')\n",
      "HLT_Photon75_R9Id... | bool                     | AsDtype('bool')\n",
      "HLT_Photon75_R9Id... | bool                     | AsDtype('bool')\n",
      "HLT_Photon90_R9Id... | bool                     | AsDtype('bool')\n",
      "HLT_Photon120_R9I... | bool                     | AsDtype('bool')\n",
      "HLT_Photon165_R9I... | bool                     | AsDtype('bool')\n",
      "HLT_Photon90_Calo... | bool                     | AsDtype('bool')\n",
      "HLT_Diphoton30_22... | bool                     | AsDtype('bool')\n",
      "HLT_Diphoton30_22... | bool                     | AsDtype('bool')\n",
      "HLT_Diphoton30PV_... | bool                     | AsDtype('bool')\n",
      "HLT_Diphoton30PV_... | bool                     | AsDtype('bool')\n",
      "HLT_Photon35_TwoP... | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu24_TwoPr... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_Jpsi_... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_Jpsi_... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_Jpsi     | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_Jpsi_... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_Jpsi_... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_Jpsi_... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_Jpsi3... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_Upsil... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_Upsil... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_Upsil... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_Upsil... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_Upsil... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_Upsil... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_Upsil... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_LowMa... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_LowMa... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_LowMass  | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_LowMa... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_LowMa... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_LowMa... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_Upsil... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon0_Upsil... | bool                     | AsDtype('bool')\n",
      "HLT_TripleMu_5_3_... | bool                     | AsDtype('bool')\n",
      "HLT_TripleMu_10_5... | bool                     | AsDtype('bool')\n",
      "HLT_TripleMu_12_10_5 | bool                     | AsDtype('bool')\n",
      "HLT_Tau3Mu_Mu7_Mu... | bool                     | AsDtype('bool')\n",
      "HLT_Tau3Mu_Mu7_Mu... | bool                     | AsDtype('bool')\n",
      "HLT_Tau3Mu_Mu7_Mu... | bool                     | AsDtype('bool')\n",
      "HLT_Tau3Mu_Mu7_Mu... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu3_DZ_... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu3_DZ_... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu3_DZ_... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu3_Trk... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu4_Jps... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu4_Jps... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu4_Jps... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu43NoF... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu48NoF... | bool                     | AsDtype('bool')\n",
      "HLT_Mu43NoFilters... | bool                     | AsDtype('bool')\n",
      "HLT_Mu48NoFilters... | bool                     | AsDtype('bool')\n",
      "HLT_Mu38NoFilters... | bool                     | AsDtype('bool')\n",
      "HLT_Mu43NoFilters... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu33NoF... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu40NoF... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu20_7_... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu20_7_... | bool                     | AsDtype('bool')\n",
      "HLT_HT425            | bool                     | AsDtype('bool')\n",
      "HLT_HT430_Displac... | bool                     | AsDtype('bool')\n",
      "HLT_HT500_Displac... | bool                     | AsDtype('bool')\n",
      "HLT_HT430_Displac... | bool                     | AsDtype('bool')\n",
      "HLT_HT400_Displac... | bool                     | AsDtype('bool')\n",
      "HLT_HT650_Displac... | bool                     | AsDtype('bool')\n",
      "HLT_HT550_Displac... | bool                     | AsDtype('bool')\n",
      "HLT_DiJet110_35_M... | bool                     | AsDtype('bool')\n",
      "HLT_DiJet110_35_M... | bool                     | AsDtype('bool')\n",
      "HLT_DiJet110_35_M... | bool                     | AsDtype('bool')\n",
      "HLT_TripleJet110_... | bool                     | AsDtype('bool')\n",
      "HLT_TripleJet110_... | bool                     | AsDtype('bool')\n",
      "HLT_TripleJet110_... | bool                     | AsDtype('bool')\n",
      "HLT_Ele30_eta2p1_... | bool                     | AsDtype('bool')\n",
      "HLT_Ele28_eta2p1_... | bool                     | AsDtype('bool')\n",
      "HLT_Ele28_HighEta... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu20_7_... | bool                     | AsDtype('bool')\n",
      "HLT_Ele15_IsoVVVL... | bool                     | AsDtype('bool')\n",
      "HLT_Ele15_IsoVVVL... | bool                     | AsDtype('bool')\n",
      "HLT_Ele15_IsoVVVL... | bool                     | AsDtype('bool')\n",
      "HLT_Ele50_IsoVVVL... | bool                     | AsDtype('bool')\n",
      "HLT_Ele15_IsoVVVL... | bool                     | AsDtype('bool')\n",
      "HLT_Mu4_TrkIsoVVL... | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_TrkIsoVVL... | bool                     | AsDtype('bool')\n",
      "HLT_Mu10_TrkIsoVV... | bool                     | AsDtype('bool')\n",
      "HLT_Mu15_IsoVVVL_... | bool                     | AsDtype('bool')\n",
      "HLT_Mu15_IsoVVVL_... | bool                     | AsDtype('bool')\n",
      "HLT_Mu15_IsoVVVL_... | bool                     | AsDtype('bool')\n",
      "HLT_Mu50_IsoVVVL_... | bool                     | AsDtype('bool')\n",
      "HLT_Mu15_IsoVVVL_... | bool                     | AsDtype('bool')\n",
      "HLT_Mu3er1p5_PFJe... | bool                     | AsDtype('bool')\n",
      "HLT_Mu3er1p5_PFJe... | bool                     | AsDtype('bool')\n",
      "HLT_Mu3er1p5_PFJe... | bool                     | AsDtype('bool')\n",
      "HLT_Mu3er1p5_PFJe... | bool                     | AsDtype('bool')\n",
      "HLT_Mu3er1p5_PFJe... | bool                     | AsDtype('bool')\n",
      "HLT_Mu3er1p5_PFJe... | bool                     | AsDtype('bool')\n",
      "HLT_Mu3er1p5_PFJe... | bool                     | AsDtype('bool')\n",
      "HLT_Mu3er1p5_PFJe... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon10_PsiP... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon20_Jpsi... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon12_Upsi... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon14_Phi_... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon18_PsiP... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon25_Jpsi    | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon18_PsiP... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon24_Upsi... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon24_Phi_... | bool                     | AsDtype('bool')\n",
      "HLT_Dimuon25_Jpsi... | bool                     | AsDtype('bool')\n",
      "HLT_DiMu4_Ele9_Ca... | bool                     | AsDtype('bool')\n",
      "HLT_DiMu9_Ele9_Ca... | bool                     | AsDtype('bool')\n",
      "HLT_DiMu9_Ele9_Ca... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleIsoMu20... | bool                     | AsDtype('bool')\n",
      "HLT_TrkMu12_Doubl... | bool                     | AsDtype('bool')\n",
      "HLT_TrkMu16_Doubl... | bool                     | AsDtype('bool')\n",
      "HLT_TrkMu17_Doubl... | bool                     | AsDtype('bool')\n",
      "HLT_Mu8              | bool                     | AsDtype('bool')\n",
      "HLT_Mu17             | bool                     | AsDtype('bool')\n",
      "HLT_Mu19             | bool                     | AsDtype('bool')\n",
      "HLT_Mu17_Photon30... | bool                     | AsDtype('bool')\n",
      "HLT_Ele8_CaloIdL_... | bool                     | AsDtype('bool')\n",
      "HLT_Ele12_CaloIdL... | bool                     | AsDtype('bool')\n",
      "HLT_Ele15_CaloIdL... | bool                     | AsDtype('bool')\n",
      "HLT_Ele23_CaloIdL... | bool                     | AsDtype('bool')\n",
      "HLT_Ele8_CaloIdM_... | bool                     | AsDtype('bool')\n",
      "HLT_Ele17_CaloIdM... | bool                     | AsDtype('bool')\n",
      "HLT_Ele23_CaloIdM... | bool                     | AsDtype('bool')\n",
      "HLT_Ele50_CaloIdV... | bool                     | AsDtype('bool')\n",
      "HLT_Ele115_CaloId... | bool                     | AsDtype('bool')\n",
      "HLT_Ele135_CaloId... | bool                     | AsDtype('bool')\n",
      "HLT_Ele145_CaloId... | bool                     | AsDtype('bool')\n",
      "HLT_Ele200_CaloId... | bool                     | AsDtype('bool')\n",
      "HLT_Ele250_CaloId... | bool                     | AsDtype('bool')\n",
      "HLT_Ele300_CaloId... | bool                     | AsDtype('bool')\n",
      "HLT_PFHT330PT30_Q... | bool                     | AsDtype('bool')\n",
      "HLT_PFHT330PT30_Q... | bool                     | AsDtype('bool')\n",
      "HLT_PFHT400_SixPF... | bool                     | AsDtype('bool')\n",
      "HLT_PFHT400_SixPF... | bool                     | AsDtype('bool')\n",
      "HLT_PFHT450_SixPF... | bool                     | AsDtype('bool')\n",
      "HLT_PFHT450_SixPF... | bool                     | AsDtype('bool')\n",
      "HLT_PFHT350          | bool                     | AsDtype('bool')\n",
      "HLT_PFHT350MinPFJ... | bool                     | AsDtype('bool')\n",
      "HLT_Photon60_R9Id... | bool                     | AsDtype('bool')\n",
      "HLT_Photon60_R9Id... | bool                     | AsDtype('bool')\n",
      "HLT_Photon60_R9Id... | bool                     | AsDtype('bool')\n",
      "HLT_ECALHT800        | bool                     | AsDtype('bool')\n",
      "HLT_DiSC30_18_EIs... | bool                     | AsDtype('bool')\n",
      "HLT_Physics          | bool                     | AsDtype('bool')\n",
      "HLT_Physics_part0    | bool                     | AsDtype('bool')\n",
      "HLT_Physics_part1    | bool                     | AsDtype('bool')\n",
      "HLT_Physics_part2    | bool                     | AsDtype('bool')\n",
      "HLT_Physics_part3    | bool                     | AsDtype('bool')\n",
      "HLT_Physics_part4    | bool                     | AsDtype('bool')\n",
      "HLT_Physics_part5    | bool                     | AsDtype('bool')\n",
      "HLT_Physics_part6    | bool                     | AsDtype('bool')\n",
      "HLT_Physics_part7    | bool                     | AsDtype('bool')\n",
      "HLT_Random           | bool                     | AsDtype('bool')\n",
      "HLT_ZeroBias         | bool                     | AsDtype('bool')\n",
      "HLT_ZeroBias_Alig... | bool                     | AsDtype('bool')\n",
      "HLT_ZeroBias_part0   | bool                     | AsDtype('bool')\n",
      "HLT_ZeroBias_part1   | bool                     | AsDtype('bool')\n",
      "HLT_ZeroBias_part2   | bool                     | AsDtype('bool')\n",
      "HLT_ZeroBias_part3   | bool                     | AsDtype('bool')\n",
      "HLT_ZeroBias_part4   | bool                     | AsDtype('bool')\n",
      "HLT_ZeroBias_part5   | bool                     | AsDtype('bool')\n",
      "HLT_ZeroBias_part6   | bool                     | AsDtype('bool')\n",
      "HLT_ZeroBias_part7   | bool                     | AsDtype('bool')\n",
      "HLT_AK4CaloJet30     | bool                     | AsDtype('bool')\n",
      "HLT_AK4CaloJet40     | bool                     | AsDtype('bool')\n",
      "HLT_AK4CaloJet50     | bool                     | AsDtype('bool')\n",
      "HLT_AK4CaloJet80     | bool                     | AsDtype('bool')\n",
      "HLT_AK4CaloJet100    | bool                     | AsDtype('bool')\n",
      "HLT_AK4CaloJet120    | bool                     | AsDtype('bool')\n",
      "HLT_AK4PFJet30       | bool                     | AsDtype('bool')\n",
      "HLT_AK4PFJet50       | bool                     | AsDtype('bool')\n",
      "HLT_AK4PFJet80       | bool                     | AsDtype('bool')\n",
      "HLT_AK4PFJet100      | bool                     | AsDtype('bool')\n",
      "HLT_AK4PFJet120      | bool                     | AsDtype('bool')\n",
      "HLT_SinglePhoton1... | bool                     | AsDtype('bool')\n",
      "HLT_SinglePhoton2... | bool                     | AsDtype('bool')\n",
      "HLT_SinglePhoton3... | bool                     | AsDtype('bool')\n",
      "HLT_Photon20_Hove... | bool                     | AsDtype('bool')\n",
      "HLT_Photon30_Hove... | bool                     | AsDtype('bool')\n",
      "HLT_EcalCalibration  | bool                     | AsDtype('bool')\n",
      "HLT_HcalCalibration  | bool                     | AsDtype('bool')\n",
      "HLT_L1UnpairedBun... | bool                     | AsDtype('bool')\n",
      "HLT_L1UnpairedBun... | bool                     | AsDtype('bool')\n",
      "HLT_L1NotBptxOR      | bool                     | AsDtype('bool')\n",
      "HLT_L1_CDC_Single... | bool                     | AsDtype('bool')\n",
      "HLT_CDC_L2cosmic_... | bool                     | AsDtype('bool')\n",
      "HLT_CDC_L2cosmic_... | bool                     | AsDtype('bool')\n",
      "HLT_HcalNZS          | bool                     | AsDtype('bool')\n",
      "HLT_HcalPhiSym       | bool                     | AsDtype('bool')\n",
      "HLT_HcalIsolatedb... | bool                     | AsDtype('bool')\n",
      "HLT_IsoTrackHB       | bool                     | AsDtype('bool')\n",
      "HLT_IsoTrackHE       | bool                     | AsDtype('bool')\n",
      "HLT_ZeroBias_Firs... | bool                     | AsDtype('bool')\n",
      "HLT_ZeroBias_Isol... | bool                     | AsDtype('bool')\n",
      "HLT_ZeroBias_Firs... | bool                     | AsDtype('bool')\n",
      "HLT_ZeroBias_Last... | bool                     | AsDtype('bool')\n",
      "HLT_ZeroBias_Firs... | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu24_eta2p... | bool                     | AsDtype('bool')\n",
      "HLT_MediumCharged... | bool                     | AsDtype('bool')\n",
      "HLT_MediumCharged... | bool                     | AsDtype('bool')\n",
      "HLT_MediumCharged... | bool                     | AsDtype('bool')\n",
      "HLT_MediumCharged... | bool                     | AsDtype('bool')\n",
      "HLT_MediumCharged... | bool                     | AsDtype('bool')\n",
      "HLT_MediumCharged... | bool                     | AsDtype('bool')\n",
      "HLT_MediumCharged... | bool                     | AsDtype('bool')\n",
      "HLT_MediumCharged... | bool                     | AsDtype('bool')\n",
      "HLT_MediumCharged... | bool                     | AsDtype('bool')\n",
      "HLT_MediumCharged... | bool                     | AsDtype('bool')\n",
      "HLT_MediumCharged... | bool                     | AsDtype('bool')\n",
      "HLT_Ele16_Ele12_E... | bool                     | AsDtype('bool')\n",
      "HLT_Rsq0p35          | bool                     | AsDtype('bool')\n",
      "HLT_Rsq0p40          | bool                     | AsDtype('bool')\n",
      "HLT_RsqMR300_Rsq0... | bool                     | AsDtype('bool')\n",
      "HLT_RsqMR320_Rsq0... | bool                     | AsDtype('bool')\n",
      "HLT_RsqMR300_Rsq0... | bool                     | AsDtype('bool')\n",
      "HLT_RsqMR320_Rsq0... | bool                     | AsDtype('bool')\n",
      "HLT_IsoMu27_MET90    | bool                     | AsDtype('bool')\n",
      "HLT_DoubleTightCh... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMediumC... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMediumC... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleTightCh... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMediumC... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleTightCh... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMediumC... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleTightCh... | bool                     | AsDtype('bool')\n",
      "HLT_VBF_DoubleLoo... | bool                     | AsDtype('bool')\n",
      "HLT_VBF_DoubleMed... | bool                     | AsDtype('bool')\n",
      "HLT_VBF_DoubleTig... | bool                     | AsDtype('bool')\n",
      "HLT_Photon50_R9Id... | bool                     | AsDtype('bool')\n",
      "HLT_Photon75_R9Id... | bool                     | AsDtype('bool')\n",
      "HLT_Photon75_R9Id... | bool                     | AsDtype('bool')\n",
      "HLT_PFMET100_PFMH... | bool                     | AsDtype('bool')\n",
      "HLT_PFMETNoMu100_... | bool                     | AsDtype('bool')\n",
      "HLT_PFMETTypeOne1... | bool                     | AsDtype('bool')\n",
      "HLT_Mu18_Mu9_Same... | bool                     | AsDtype('bool')\n",
      "HLT_Mu18_Mu9_Same... | bool                     | AsDtype('bool')\n",
      "HLT_Mu18_Mu9         | bool                     | AsDtype('bool')\n",
      "HLT_Mu18_Mu9_DZ      | bool                     | AsDtype('bool')\n",
      "HLT_Mu20_Mu10_Sam... | bool                     | AsDtype('bool')\n",
      "HLT_Mu20_Mu10_Sam... | bool                     | AsDtype('bool')\n",
      "HLT_Mu20_Mu10        | bool                     | AsDtype('bool')\n",
      "HLT_Mu20_Mu10_DZ     | bool                     | AsDtype('bool')\n",
      "HLT_Mu23_Mu12_Sam... | bool                     | AsDtype('bool')\n",
      "HLT_Mu23_Mu12_Sam... | bool                     | AsDtype('bool')\n",
      "HLT_Mu23_Mu12        | bool                     | AsDtype('bool')\n",
      "HLT_Mu23_Mu12_DZ     | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu2_Jps... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu2_Jps... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleMu3_DCA... | bool                     | AsDtype('bool')\n",
      "HLT_TripleMu_5_3_... | bool                     | AsDtype('bool')\n",
      "HLT_QuadPFJet98_8... | bool                     | AsDtype('bool')\n",
      "HLT_QuadPFJet103_... | bool                     | AsDtype('bool')\n",
      "HLT_QuadPFJet111_... | bool                     | AsDtype('bool')\n",
      "HLT_QuadPFJet98_8... | bool                     | AsDtype('bool')\n",
      "HLT_QuadPFJet103_... | bool                     | AsDtype('bool')\n",
      "HLT_QuadPFJet105_... | bool                     | AsDtype('bool')\n",
      "HLT_QuadPFJet111_... | bool                     | AsDtype('bool')\n",
      "HLT_QuadPFJet98_8... | bool                     | AsDtype('bool')\n",
      "HLT_QuadPFJet103_... | bool                     | AsDtype('bool')\n",
      "HLT_QuadPFJet105_... | bool                     | AsDtype('bool')\n",
      "HLT_QuadPFJet111_... | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet330_T... | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet330_T... | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet330_T... | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet330_T... | bool                     | AsDtype('bool')\n",
      "HLT_AK8PFJet330_T... | bool                     | AsDtype('bool')\n",
      "HLT_Diphoton30_18... | bool                     | AsDtype('bool')\n",
      "HLT_Diphoton30_18... | bool                     | AsDtype('bool')\n",
      "HLT_Mu12_IP6_part0   | bool                     | AsDtype('bool')\n",
      "HLT_Mu12_IP6_part1   | bool                     | AsDtype('bool')\n",
      "HLT_Mu12_IP6_part2   | bool                     | AsDtype('bool')\n",
      "HLT_Mu12_IP6_part3   | bool                     | AsDtype('bool')\n",
      "HLT_Mu12_IP6_part4   | bool                     | AsDtype('bool')\n",
      "HLT_Mu9_IP5_part0    | bool                     | AsDtype('bool')\n",
      "HLT_Mu9_IP5_part1    | bool                     | AsDtype('bool')\n",
      "HLT_Mu9_IP5_part2    | bool                     | AsDtype('bool')\n",
      "HLT_Mu9_IP5_part3    | bool                     | AsDtype('bool')\n",
      "HLT_Mu9_IP5_part4    | bool                     | AsDtype('bool')\n",
      "HLT_Mu7_IP4_part0    | bool                     | AsDtype('bool')\n",
      "HLT_Mu7_IP4_part1    | bool                     | AsDtype('bool')\n",
      "HLT_Mu7_IP4_part2    | bool                     | AsDtype('bool')\n",
      "HLT_Mu7_IP4_part3    | bool                     | AsDtype('bool')\n",
      "HLT_Mu7_IP4_part4    | bool                     | AsDtype('bool')\n",
      "HLT_Mu9_IP4_part0    | bool                     | AsDtype('bool')\n",
      "HLT_Mu9_IP4_part1    | bool                     | AsDtype('bool')\n",
      "HLT_Mu9_IP4_part2    | bool                     | AsDtype('bool')\n",
      "HLT_Mu9_IP4_part3    | bool                     | AsDtype('bool')\n",
      "HLT_Mu9_IP4_part4    | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_IP5_part0    | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_IP5_part1    | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_IP5_part2    | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_IP5_part3    | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_IP5_part4    | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_IP6_part0    | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_IP6_part1    | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_IP6_part2    | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_IP6_part3    | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_IP6_part4    | bool                     | AsDtype('bool')\n",
      "HLT_Mu9_IP6_part0    | bool                     | AsDtype('bool')\n",
      "HLT_Mu9_IP6_part1    | bool                     | AsDtype('bool')\n",
      "HLT_Mu9_IP6_part2    | bool                     | AsDtype('bool')\n",
      "HLT_Mu9_IP6_part3    | bool                     | AsDtype('bool')\n",
      "HLT_Mu9_IP6_part4    | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_IP3_part0    | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_IP3_part1    | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_IP3_part2    | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_IP3_part3    | bool                     | AsDtype('bool')\n",
      "HLT_Mu8_IP3_part4    | bool                     | AsDtype('bool')\n",
      "HLT_QuadPFJet105_... | bool                     | AsDtype('bool')\n",
      "HLT_TrkMu6NoFilte... | bool                     | AsDtype('bool')\n",
      "HLT_TrkMu16NoFilt... | bool                     | AsDtype('bool')\n",
      "HLT_DoubleTrkMu_1... | bool                     | AsDtype('bool')\n",
      "HLTriggerFinalPath   | bool                     | AsDtype('bool')\n",
      "L1simulation_step    | bool                     | AsDtype('bool')\n"
     ]
    }
   ],
   "source": [
    "tree = file['Events']\n",
    "tree.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Wow, that's a lot of stuff!  What is shown here is each column (a ROOT [TBranch](https://root.cern.ch/doc/master/classTBranch.html)) inside the tree, along with some interpretation information.  Another source of this information is [this website](https://cms-nanoaod-integration.web.cern.ch/integration/master-102X/mc102X_doc.html) (and its' parent folder contains branch lists for other NanoAOD versions.)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "from coffea.nanoevents import NanoEventsFactory\n",
    "\n",
    "dyevents = NanoEventsFactory.from_root(dy_files[0]).events()\n",
    "gevents = NanoEventsFactory.from_root(gjets_files[0]).events()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['dEscaleDown',\n",
       " 'dEscaleUp',\n",
       " 'dEsigmaDown',\n",
       " 'dEsigmaUp',\n",
       " 'deltaEtaSC',\n",
       " 'dr03EcalRecHitSumEt',\n",
       " 'dr03HcalDepth1TowerSumEt',\n",
       " 'dr03TkSumPt',\n",
       " 'dr03TkSumPtHEEP',\n",
       " 'dxy',\n",
       " 'dxyErr',\n",
       " 'dz',\n",
       " 'dzErr',\n",
       " 'eCorr',\n",
       " 'eInvMinusPInv',\n",
       " 'energyErr',\n",
       " 'eta',\n",
       " 'hoe',\n",
       " 'ip3d',\n",
       " 'jetPtRelv2',\n",
       " 'jetRelIso',\n",
       " 'mass',\n",
       " 'miniPFRelIso_all',\n",
       " 'miniPFRelIso_chg',\n",
       " 'mvaFall17V2Iso',\n",
       " 'mvaFall17V2noIso',\n",
       " 'pfRelIso03_all',\n",
       " 'pfRelIso03_chg',\n",
       " 'phi',\n",
       " 'pt',\n",
       " 'r9',\n",
       " 'scEtOverPt',\n",
       " 'sieie',\n",
       " 'sip3d',\n",
       " 'mvaTTH',\n",
       " 'charge',\n",
       " 'cutBased',\n",
       " 'jetIdx',\n",
       " 'pdgId',\n",
       " 'photonIdx',\n",
       " 'tightCharge',\n",
       " 'vidNestedWPBitmap',\n",
       " 'vidNestedWPBitmapHEEP',\n",
       " 'convVeto',\n",
       " 'cutBased_HEEP',\n",
       " 'isPFcand',\n",
       " 'jetNDauCharged',\n",
       " 'lostHits',\n",
       " 'mvaFall17V2Iso_WP80',\n",
       " 'mvaFall17V2Iso_WP90',\n",
       " 'mvaFall17V2Iso_WPL',\n",
       " 'mvaFall17V2noIso_WP80',\n",
       " 'mvaFall17V2noIso_WP90',\n",
       " 'mvaFall17V2noIso_WPL',\n",
       " 'seedGain',\n",
       " 'genPartIdx',\n",
       " 'genPartFlav',\n",
       " 'cleanmask',\n",
       " 'genPartIdxG',\n",
       " 'jetIdxG',\n",
       " 'photonIdxG']"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dyevents.Electron.fields"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Ok, slightly more manageable content. Try replacing the above cell with `dyevents.Electron.fields` and `dyevents.Photon.fields` to see what EGamma-related columns exist in the tree.\n",
    "\n",
    "Below shows how to see a short description of a column (also available via the `shift`+`tab` keystroke inside jupyter editor). Try `dyevents.Electron.jetIdx.__doc__` and `dyevents.Electron.lostHits.__doc__`. **Q: What could these values be used for? Why are they associated with an electron?**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'cut-based ID Fall17 V2 (0:fail, 1:veto, 2:loose, 3:medium, 4:tight)'"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dyevents.Electron.cutBased.__doc__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Array programming primer\n",
    "\n",
    "Before diving into the content, let's give a quick overview of [awkward](https://awkward-array.org/quickstart.html) array syntax, as the NanoEvents structure is an awkward array.  It is an extension of the [numpy](https://numpy.org/) array programming syntax.  If you are unfamiliar with numpy, perhaps glance through the [numpy quickstart](https://docs.scipy.org/doc/numpy/user/quickstart.html) before continuing, in particular the universal function and boolean indexing sections.  The main extensions to numpy that are used here are jagged arrays.  Some basic properties are shown below."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<Array [[1, 2, 3], [4, -2, 4], [], [5]] type='4 * var * int64'>\n",
      "<Array [3, 3, 0, 1] type='4 * int64'>\n"
     ]
    }
   ],
   "source": [
    "import awkward as ak\n",
    "import numpy as np\n",
    "\n",
    "a = ak.Array([[1, 2, 3], [4, -2, 4], [], [5]])\n",
    "print(repr(a))\n",
    "print(repr(ak.num(a)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Above is a singly-jagged array, where the first dimension has length 4, and the second dimension has variable length.  The length along that dimension is accessible via the [ak.num function](https://awkward-array.readthedocs.io/en/latest/_auto/ak.num.html) (the default axis is 1 for `ak.num`).  This is like the array `dyevents.Electrons`.  All comparison operators and numpy ufuncs are _broadcast_ over the content, meaning they are applied to each element in the jagged array. For example:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[False, False, True], [True, False, True], [], [True]]\n",
      "[[1, 2, 3], [4, 2, 4], [], [5]]\n"
     ]
    }
   ],
   "source": [
    "print(a > 2)\n",
    "print(abs(a))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The jagged array `a > 2` can be used to filter (i.e. select) only those elements that pass the criteria. This is done using the `array[filter]` notation. We will use this syntax later to, for example, select only electrons above a certain pT threshold:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[False, False, True], [True, False, True], [], [True]]\n",
      "[[3], [4, 4], [], [5]]\n"
     ]
    }
   ],
   "source": [
    "object_cut = (a > 2)\n",
    "print(object_cut)\n",
    "print(a[object_cut])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "When using jagged arrays in a typical physics analysis, in addition to `object_cuts` we can also have event-level selections. For example, we might want to only select _events_ where `all` electrons in the event have pT above some threshold:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[False, False, True, True]\n",
      "[[], [5]]\n"
     ]
    }
   ],
   "source": [
    "event_cut = ak.all(a > 2, axis=1)\n",
    "print(event_cut)\n",
    "print(a[event_cut])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Most utilities in the scientific python \"speak\" numpy arrays, however they do not know about jagged arrays.  So, for example, when interfacing with [pandas](https://pandas.pydata.org/) or [matplotlib](https://matplotlib.org/) we need to remove the jaggedness by using `ak.flatten`.  Of course, we should do this only after making any event-level selections, as after flattening, that structure is lost."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[], [5]]\n",
      "[5]\n"
     ]
    }
   ],
   "source": [
    "print(a[event_cut])\n",
    "print(ak.flatten(a[event_cut]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Another common awkward type is masked arrays.  The main feature is that certain elements can be [null](https://en.wikipedia.org/wiki/Nullable_type) (syn. N/A, not available, `None`)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Array [1, 2, None, 4] type='4 * ?int64'>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "b = ak.Array([1, 2, None, 4])\n",
    "b"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The null value propagates through any operation, until it is filled in"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1, 2, None, 4]\n",
      "[1, 2, -1, 4]\n",
      "[True, True, None, True]\n",
      "[True, True, False, True]\n"
     ]
    }
   ],
   "source": [
    "print(b)\n",
    "print(ak.fill_none(b, -1))\n",
    "print(b > 0)\n",
    "print(ak.fill_none(b, -1) > 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the scientific python ecosystem, [matplotlib](https://matplotlib.org/) is the de facto standard plotting utility. Let's import it, along with [mplhep](https://mplhep.readthedocs.io/en/latest/), a helper library to make HEP plots easier (including CMS style!)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import mplhep\n",
    "mplhep.style.use(mplhep.style.CMS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# run this again to make sure the figure size updates (a known bug)\n",
    "mplhep.style.use(mplhep.style.CMS)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Kinematics\n",
    "Below is a plot of all reconstructed electrons' transverse momentum ($p_{T}$) in this NanoAOD file.\n",
    "Try plotting instead the $\\eta$ and $\\phi$ of the electrons.  Plot the electron `counts`.  Plot the $p_{T}$ of the electrons after applying a cut on $|\\eta| < 1.4$."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "ax.set_xlabel(r'Electron $p_{T}$ [GeV]')\n",
    "ax.hist(ak.flatten(dyevents.Electron.pt), bins=np.linspace(0, 200, 50));"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Below we plot the invariant mass of all distinct pairs of electrons. The pairs made using the `combinations` function. The `lep1` and `lep2` names are names we assign so we can separately access the first and second electron in the pair."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAm8AAAJmCAYAAAAKOWdAAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy81sbWrAAAACXBIWXMAAAsTAAALEwEAmpwYAAA1AUlEQVR4nO3dfZgkdXnv//cdFkERFFDwAXUBn0VF2Cii0ai/oMmmEwjmGIwJ5iR2cjSrJzGJG5KoUTSrMZcKGs14Eh/4CUaMHOmsijHRaGSBrIoYnyLCiiAqrisC8qB4nz+q2m17unt6dqa751v9fl1XXTVTdVf1t7tqZj7zrafITCRJklSGn5p1AyRJkjQ+w5skSVJBDG+SJEkFMbxJkiQVxPAmSZJUkHWzbsC0RISX1UqSpKJkZvRPs+dNkiSpIHMX3jJzosOxxx478dd4znOeM/HXmNbrNOm9uO3n8zWatO3dv9z2TXidpvy+H2XuwpskSVLJ5uact652u71oWqvVotVqzaA1kiRp3nU6HTqdztj1cxfeFhYWZt0ESZKkHxvUifSWt7xlaL2HTSVJkgpieCvQtA7xTuN1mvRepqFJn1dTXmNamvR5Nem9TEOTPq8mvZdZiqWuaGiK7n3eJv1+N2zYwPbt2yf6Glqb3Pbzy20/v9z282ka2z2iur1bep83SZKkshneJEmSCmJ4kyRJKojhTZIkqSCGt1U26CbAmg9u+/nltp9fbvv5NOvt7tWmkiRJa4xXm0qSJDWE4U2SJKkghjdJkqSCGN4kSZIKYniTJEkqiOFNkiSpIIY3SZKkghjeJEmSCmJ4kyRJKsi6WTdg2gY90qLVatFqtWbQGkmSNO86nQ6dTmfseh+PJUmStMb4eCxJkqSGmLvDppLmx/rNW8eq27Fl44RbIkmrx543SZKkghjeJEmSCmJ4kyRJKojhTZIkqSCGN0mSpIIY3iRJkgpieJMkSSqI4U2SJKkghjdJkqSCGN4kSZIKYniTJEkqiOFNkiSpIIY3SZKkghjeJEmSCmJ4kyRJKsi6WTdg2trt9qJprVaLVqs1g9ZIkqR51+l06HQ6Y9dHZk6wOWtHRCTAvLxfSbB+89ax6nZs2TjhlkjS8kQEAJkZ/fM8bCpJklQQw5skSVJBDG+SJEkFMbxJkiQVxPAmSZJUEMObJElSQQxvkiRJBTG8SZIkFcTwJkmSVJCxw1tEvCwiPjBieFhf/ckRsS0iboyInRFxfkQcPWL9E62XJElqguU82/Qk4KgR80/vfhERLwBeV3/7OeCuQAs4ISKenJkX9i446XpJkqSmGKvnLaoHbB0JfCYzY8jwibr2IOBVwM3AYzPzqMw8DHg+sA9wZt+6J1ovSZLUJOMeNr0ncEfg8jFqT6EKUadn5kXdiZl5JnABcExEHDXFekmSpMYYN7zdvx5/eYzaU+rxeQPmnddXM416SZKkxhj3nLduePtqRLSB44G9gM8A52bmV3tqDweuz8wvDFhP91y0I6ZYL0mS1Bjjhrcj6/FrgP16pj8LeHFEPC8zz6rPjTsEuGrIenbW40Phx+fSTaxekiSpaZZ72HQnsBE4iKp368XAvsDfR8RDgYOpAuGuIevpD1eTrpckSWqUcXvezgc+BZyVmV+vp+0CXh4RtwOvAF4CPG+J9exVj/ce83VXvX7Dhg1jrgra7TbtdnvsekmSpGEWFhZYWFhY8XoiM1e2goi7UgW5LwMPAm4DvpaZi847i4jDgK8BH8vMJ9aHQSdW3zcvAVb6fiWVY/3mrWPV7diyccItkaTlqSIPZGb0z1vx47Ey87vAt4H7ZpWMrqM6vDlId/q19bITrZckSWqaJcNbRNw9Ik6NiCcPmb8OuAvwlXrSFcABQ+61dnw9vrJn2qTrJUmSGmOcnrcbgTcC50bEHQfMfwrVOWaX1t+fU49PGlB7Yl/NNOolSZIaY8nwlpk3A/9EdYXpWyPizt159YPg/xb4AfDKevLZwC3AaRFxXE/tJuAE4JLMvKznJSZdL0mS1BjjnvP2v6kOiz4DuCoiPh4Rnwe2A/cF/igzPweQmbuAzVS3ENkWEZdFxNXAGVTPI93Uu+JJ10uSJDXJWOGtDkyPpnog/DeAY6ieL/o+4PjMPKOv/vXA04GLqW7wux/V7UYek5mXDFj/ROslSZKaYsW3CimFtwqR5o+3CpFUqoneKkSSJEnTY3iTJEkqiOFNkiSpIIY3SZKkghjeJEmSCmJ4kyRJKojhTZIkqSCGN0mSpIIY3iRJkgpieJMkSSqI4U2SJKkg62bdgGlrt9uLprVaLVqt1gxaI0mS5l2n06HT6Yxd74PpJTWWD6aXVCofTC9JktQQhjdJkqSCGN4kSZIKYniTJEkqiOFNkiSpIIY3SZKkghjeJEmSCmJ4kyRJKojhTZIkqSCGN0mSpIIY3iRJkgpieJMkSSqI4U2SJKkghjdJkqSCGN4kSZIKYniTJEkqiOFNkiSpIOtm3YBpa7fbi6a1Wi1ardYMWiNJkuZdp9Oh0+mMXR+ZOcHmrB0RkQDz8n4lwfrNW8eq27Fl44RbIknLExEAZGb0z/OwqSRJUkEMb5IkSQUxvEmSJBXE8CZJklQQw5skSVJBDG+SJEkFMbxJkiQVxPAmSZJUEMObJElSQQxvkiRJBTG8SZIkFcTwJkmSVBDDmyRJUkEMb5IkSQUxvEmSJBXE8CZJklQQw5skSVJB1s26AdPWbrcXTWu1WrRarRm0RpIkzbtOp0On0xm7PjJzgs1ZOyIiAebl/UqC9Zu3jlW3Y8vGCbdEkpYnIgDIzOif52FTSZKkghjeJEmSCmJ4kyRJKojhTZIkqSCGN0mSpIIY3iRJkgpieJMkSSqI4U2SJKkghjdJkqSCGN4kSZIKYniTJEkqiOFNkiSpIIY3SZKkghjeJEmSCmJ4kyRJKojhTZIkqSCGN0mSpIIY3iRJkgqybtYNmLZ2u71oWqvVotVqzaA1kiRp3nU6HTqdztj1kZkTbM7aEREJMC/vVxKs37x1rLodWzZOuCWStDwRAUBmRv88D5tKkiQVxPAmSZJUEMObJElSQQxvkiRJBTG8SZIkFcTwJkmSVBDDmyRJUkEMb5IkSQUxvEmSJBVkj8NbVC6IiIyIvQbMPzkitkXEjRGxMyLOj4ijR6xvovWSJElNsJKet98HThg0IyJeALwHOA7YAdwMtICLIuL4addLkiQ1xR6Ft4h4KPDqIfMOAl5FFagem5lHZeZhwPOBfYAzp1kvSZLUJMsObxFxB+CdwE3ArgElp1CFqNMz86LuxMw8E7gAOCYijppivSRJUmOs24NlXg4cDfwq8NfAgX3zT6nH5w1Y9jzgqXXNn02pXlKDrN+8ddZNkKSZWlbPW0Q8Efgj4B2Z+Z4hZYcD12fmFwbMu7AeHzHFekmSpMYYO7xFxF2AdwBfAzYNqQngEGDnkNV0px86jXpJkqSmWc5h0zcChwFPyszvDak5uF7noHPhYHG4mnS9JElSo4wV3iLiGcCvA6/OzI+t4PW694Pbe1b1GzZsGHNV0G63abfbY9dLkiQNs7CwwMLCworXs2R4i4jDgDcBnwH+YonyncAPgYOGzO9Ov3ZK9Yts37592CxJkqSJWU6nUHWm2GDj9Lw9heqK0q8D7+tbWffw5Psj4kfAy4DrqA5vDtKdfi1AZmZETKxekiSpaZZztenDgKf1DXes551Qf38IcAVwwJB7rXWffnBlz7RJ10uSJDXGkuEtM9+emTFooHo0FcC6etr7gHPqaScNWN2J9ficnmmTrpckSWqMlTzbdJizgVuA0yLiuO7EiNhE1UN3SWZeNsV6SZKkxlj18JaZu4DNwL7Atoi4LCKuBs6geh7ppmnWS5IkNckket7IzNcDTwcuBo4E9gPOBx6TmZdMu16SJKkpIjNn3YapiIgEmJf3KzXVJJ5tumPLxlVfpyStRPfuHvU1Bj9hIj1vkiRJmgzDmyRJUkEMb5IkSQUxvEmSJBXE8CZJklQQw5skSVJBDG+SJEkFMbxJkiQVxPAmSZJUEMObJElSQdbNugHT1m63F01rtVq0Wq0ZtEaSJM27TqdDp9MZu95nm0oqis82lTQPfLapJElSQxjeJEmSCmJ4kyRJKojhTZIkqSCGN0mSpIIY3iRJkgpieJMkSSqI4U2SJKkghjdJkqSCGN4kSZIKYniTJEkqiOFNkiSpIIY3SZKkghjeJEmSCmJ4kyRJKojhTZIkqSCGN0mSpIKsm3UDpq3dbi+a1mq1aLVaM2iNJEmad51Oh06nM3Z9ZOYEm7N2REQCzMv7lZpq/eatq77OHVs2rvo6JWklIgKAzIz+eR42lSRJKojhTZIkqSCGN0mSpIIY3iRJkgpieJMkSSqI4U2SJKkghjdJkqSCGN4kSZIKYniTJEkqiOFNkiSpIIY3SZKkghjeJEmSCmJ4kyRJKojhTZIkqSCGN0mSpIIY3iRJkgpieJMkSSqI4U2SJKkg62bdgGlrt9uLprVaLVqt1gxaI0mS5l2n06HT6YxdH5k5weasHRGRAPPyfqWmWr9566qvc8eWjau+TklaiYgAIDOjf56HTSVJkgpieJMkSSqI4U2SJKkghjdJkqSCGN4kSZIKYniTJEkqiOFNkiSpIIY3SZKkghjeJEmSCmJ4kyRJKojhTZIkqSCGN0mSpIIY3iRJkgpieJMkSSqI4U2SJKkghjdJkqSCGN4kSZIKsm7WDZi2dru9aFqr1aLVas2gNZIkad51Oh06nc7Y9ZGZE2zO2hERCTAv71dqqvWbt676Onds2bjq65SklYgIADIz+ud52FSSJKkghjdJkqSCGN4kSZIKYniTJEkqiOFNkiSpIIY3SZKkghjeJEmSCmJ4kyRJKojhTZIkqSBjh7eI2DciTo+I/4yIGyLiyog4PyI2DKk/OSK2RcSNEbGzrj16xPonWi9JktQEY4W3iLgDcDHwZ8ADgc8CtwIt4OKIeEZf/QuA9wDHATuAm+vaiyLi+AHrn2i9JElSU4zb8/Y84BHAvwD3yczjM/PBwMlAAG+IiP0AIuIg4FVUgeqxmXlUZh4GPB/YBzizd8WTrpckSWqSccPb0+vxH2Tm97oTM/O9wD8DdwMeXk8+hSpEnZ6ZF/XUnglcABwTEUf1rHvS9ZIkSY0xbng7HLg1Mz83YN4X6/ER9fiUenzegNrz+mqmUS9JktQY68as+zXgB0PmHVmPr67HhwPXZ+YXBtReWI+P6Jk26XpJkqTGGCu8ZebHer+PiP2BewK/DvwKcCnwHxERwCHAVUNWtbMeH1qvZ6L1kiRJTTNuz9uPRcRTgA/3TPoY8KuZ+aOIuFu9zl1DFu8PVwdPuF6SJKlRlh3eqALSBcA9gAcDjwdOj4jnjrHsXvV47zFfa9XrN2wYeFu6gdrtNu12e+x6SZKkYRYWFlhYWFjxepYd3jLzUuBpABFxD+Ac4DlU91v7K+CHwEFDFu9Ov7Ye75xw/SLbt28fNkuSJGliltMpVJ0pNtiKHo+Vmd8Afr/+9qTMTOA6qsObg3SnX1svP9F6SZKkplkyvEXEcRHx5Yg4Y0hJ9+KBO9XjK4ADhtxrrfv0gyt7pk26XpIkqTHG6Xn7b+D+wP8Xg/vwfroeX1aPz6nHJw2oPbGvZhr1kiRJjbFkeMvM71BdUfoQ4C8iontRABHxIOAN9bf/WI/PBm4BTouI43pqNwEnAJdkZjfoTaNekiSpMcY95+13gZuAvwSuiYiPRcRlwH9Rhbo3Zeb/BcjMXcBmYF9gW0RcFhFXA2dQPY90U++KJ10vSZLUJGOFt8z8ItVtQf4e+A5wLHAA1f3efiEzn9tX/3qq56FeTPUEhv2A84HHZOYlA9Y/0XpJkqSmiOoCzuaLiASYl/crNdX6zVtXfZ07tmxc9XVK0kp0LzPIzEXXG6zoViGSJEmaLsObJElSQQxvkiRJBTG8SZIkFcTwJkmSVBDDmyRJUkEMb5IkSQUxvEmSJBXE8CZJklQQw5skSVJBDG+SJEkFWTfrBkxbu91eNK3VatFqtWbQGkmSNO86nQ6dTmfseh9ML6koPphe0jzwwfSSJEkNYXiTJEkqiOFNkiSpIIY3SZKkghjeJEmSCmJ4kyRJKojhTZIkqSCGN0mSpIIY3iRJkgpieJMkSSqI4U2SJKkghjdJkqSCGN4kSZIKYniTJEkqiOFNkiSpIIY3SZKkghjeJEmSCrJu1g2Ytna7vWhaq9Wi1WrNoDWSJGnedTodOp3O2PWRmRNsztoREQkwL+9Xaqr1m7eu+jp3bNm46uuUpJWICAAyM/rnedhUkiSpIIY3SZKkghjeJEmSCmJ4kyRJKojhTZIkqSCGN0mSpIIY3iRJkgpieJMkSSqI4U2SJKkghjdJkqSCGN4kSZIKYniTJEkqiOFNkiSpIIY3SZKkghjeJEmSCmJ4kyRJKojhTZIkqSDrZt2AaWu324umtVotWq3WDFojSZLmXafTodPpjF0fmTnB5qwdEZEA8/J+paZav3nrqq9zx5aNq75OSVqJiAAgM6N/nodNJUmSCmJ4kyRJKojhTZIkqSCGN0mSpIIY3iRJkgpieJMkSSqI4U2SJKkghjdJkqSCGN4kSZIKYniTJEkqiOFNkiSpIIY3SZKkghjeJEmSCmJ4kyRJKojhTZIkqSCGN0mSpIIY3iRJkgpieJMkSSrIulk3YNra7faiaa1Wi1arNYPWSJKkedfpdOh0OmPXR2ZOsDlrR0QkwLy8X6mp1m/euurr3LFl46qvU5JWIiIAyMzon+dhU0mSpIIY3iRJkgpieJMkSSqI4U2SJKkghjdJkqSCGN4kSZIKYniTJEkqiOFNkiSpIIY3SZKkgiwrvEXEMyPiIxFxbUR8OyL+NSJOHVJ7ckRsi4gbI2JnRJwfEUePWPdE6yVJkppg7PAWEW8B3gk8DvgmcBXwM8DbIuK90X2OQ1X7AuA9wHHADuBmoAVcFBHHD1j3ROslSZKaYqzwFhHHAb9DFdgemZlHZ+YxwEOBzwInAe269iDgVVSB6rGZeVRmHgY8H9gHOLNv3ROtlyRJapJxe96eXY//MjO/0J2YmZcD3cOmv1mPT6EKUadn5kU9tWcCFwDHRMRRPeuedL0kSVJjjBvejqjHH+2fkZmfBnYBj6gnnVKPzxuwnvP6aqZRL0mS1Bjrxqy7CPgqcE3/jIjYF7gzsLOedDhwfW8PXY8L6/ERPdMmXS9JktQYY4W3zHzxiNkvAvYGPlZftHAI1blxg3QD3qEAk66XJElqmnF73hapg9TpwGnArcArgYPrde4aslh/uJp0vSRJUqPsUXiLiGOBN1DdquM24NTM/ExE3G2JRfeqx3uP+VKrXr9hw4YxVwXtdpt2uz12vSRJ0jALCwssLCyseD3LCm8RsQ/wcuCFVBc7fJYquH26LtkJ/BA4aMgqutOvnVL9Itu3bx82S5IkaWKW0ynUc/vcRcYObxFxb+CDwFFUIeovgIXMvL1bk5kZEddRHd4cpDv92mnUS5IkNc24N+ndH9hKFdwuBh6RmW/qDW49rgAOGHKvte7TD66cYr0kSVJjjHuft+cDjwTeCzwxM78+ovacenzSgHkn9tVMo16SJKkxxg1vzwFuB34vM29dovZs4BbgtPqxWgBExCbgBOCSzLxsivWSJEmNseQ5bxFxMHA/qsD0jhEn0O3KzGdm5q6I2Ay8DtgWEZ+lupDg3lTPI93Uu9Ck6yVJkppknAsW1tfjfYGnjaj7RveLzHx9RFwN/DHwcKrbiZwP/HlmfrZ/wUnXS5IkNUVk5qzbMBURkQDz8n6lplq/eeuqr3PHlo2rvk5JWonukc7MXHTIc9xz3iRJkrQGGN4kSZIKYniTJEkqiOFNkiSpIIY3SZKkghjeJEmSCmJ4kyRJKojhTZIkqSCGN0mSpIIY3iRJkgoyzrNNG6Xdbi+a1mq1aLVaM2iNJEmad51Oh06nM3a9zzaVVBSfbSppHvhsU0mSpIYwvEmSJBXE8CZJklQQw5skSVJBDG+SJEkFMbxJkiQVxPAmSZJUEMObJElSQQxvkiRJBTG8SZIkFcTwJkmSVBDDmyRJUkEMb5IkSQUxvEmSJBXE8CZJklQQw5skSVJBDG+SJEkFWTfrBkxbu91eNK3VatFqtWbQGkmSNO86nQ6dTmfs+sjMCTZn7YiIBJiX9ys11frNW1d9nTu2bFz1dUrSSkQEAJkZ/fM8bCpJklQQw5skSVJBDG+SJEkFMbxJkiQVxPAmSZJUEMObJElSQQxvkiRJBTG8SZIkFWTunrAgaW2axM13JamJ7HmTJEkqiOFNkiSpIIY3SZKkgnjOm6S5t5zz7XyIvaRZs+dNkiSpIIY3SZKkghjeJEmSCmJ4kyRJKojhTZIkqSBzd7Vpu91eNK3VatFqtWbQGkmSNO86nQ6dTmfs+sjMCTZn7YiIBJiX9yuVppTHY3mrEEnTEBEAZGb0z/OwqSRJUkEMb5IkSQUxvEmSJBXE8CZJklQQw5skSVJBDG+SJEkFMbxJkiQVxPAmSZJUEMObJElSQQxvkiRJBTG8SZIkFcTwJkmSVBDDmyRJUkEMb5IkSQUxvEmSJBXE8CZJklQQw5skSVJB1s26AdPWbrcXTWu1WrRarRm0RpIkzbtOp0On0xm7PjJzgs1ZOyIiAebl/UqlWb9566ybMJYdWzbOugmS5kBEAJCZ0T/Pw6aSJEkFMbxJkiQVxPAmSZJUEMObJElSQQxvkiRJBTG8SZIkFcTwJkmSVBDDmyRJUkEMb5IkSQVZdniLyjUR8fIRNSdHxLaIuDEidkbE+RFx9KzqJUmSmmJPet5+AbjXsJkR8QLgPcBxwA7gZqAFXBQRx0+7XpIkqUnGDm8RsX9EPAv4hxE1BwGvogpUj83MozLzMOD5wD7AmdOslyRJapqxwltEnAtcD5wFHDKi9BSqEHV6Zl7UnZiZZwIXAMdExFFTrJckSWqUcXveLgT+Dngz8JERdafU4/MGzDuvr2Ya9ZIkSY2ybpyizHxt9+uIOBV40pDSw4HrM/MLA+ZdWI+PmGK9JElSo6zarUIiIqgOqe4cUtKdfug06iVJkppoNe/zdjBVT96uIfP7w9Wk6yVJkhpnrMOmq2Sverz3LOs3bNgw5uqg3W7TbrfHrpckSRpmYWGBhYWFFa9nNcPbTuCHwEFD5nenXzul+oG2b98+arYkSdJELKdTqDpbbLBVO2yamQlcR3V4c5Du9GunUS9JktREq/1s0yuAA4bca6379IMrp1gvSZLUKKsd3s6pxycNmHdiX8006iVJkhpltcPb2cAtwGkRcVx3YkRsAk4ALsnMy6ZYL0mS1CirGt4ycxewGdgX2BYRl0XE1cAZVM8j3TTNekmSpKZZ7Z43MvP1wNOBi4Ejgf2A84HHZOYl066XJElqkqgu4my+iEiAeXm/UmnWb9466yaMZceWjbNugqQ50L1VSGYuumfIqve8SZIkaXIMb5IkSQUxvEmSJBXE8CZJklQQw5skSVJBDG+SJEkFMbxJkiQVxPAmSZJUEMObJElSQQxvkiRJBTG8SZIkFWTdrBswbe12e9G0VqtFq9WaQWskSdK863Q6dDqdset9ML2kNcEH00vSbj6YXpIkqSEMb5IkSQUxvEmSJBXE8CZJklQQw5skSVJB5u5WIZKmq5SrSCWpFPa8SZIkFcTwJkmSVBDDmyRJUkEMb5IkSQUxvEmSJBXE8CZJklQQw5skSVJBDG+SJEkFMbxJkiQVxPAmSZJUEMObJElSQebu2abtdnvRtFarRavVmkFrJEnSvOt0OnQ6nbHrIzMn2Jy1IyISYF7er7RWNO3B9Du2bJx1EyTNgYgAIDOjf56HTSVJkgpieJMkSSqI4U2SJKkghjdJkqSCGN4kSZIKMne3CtHsjXv1oVf1SZK0mOFNq6Jpt4OQJGmtMrxpzbKHTpKkxQxvGskeNUmS1hYvWJAkSSqI4U2SJKkgHjaVtGweTpek2bHnTZIkqSD2vKl4y+kF8spUSVLp7HmTJEkqiD1vc8jzlVaPvX6SpGmz502SJKkgc9fz1m63F01rtVq0Wq0ZtEaSJM27TqdDp9MZuz4yc4LNWTsiIgGa/H49HLq01T50OYnPvITDq/O8r5WwfSSVLyIAyMzonzd3PW+abz4vVZJUOsObpB+b5x61cfkPgKRZM7xJAxhiJElrlVebSpIkFcSeN2mN8bCcJGkUe94kSZIKYs+bVCjPy5Ok+WTPmyRJUkEMb5IkSQUxvEmSJBXEc94K4LlNkiSpy/C2yrzNgyRJmiQPm0qSJBXE8CZJklQQw5skSVJBDG+SJEkFMbxJkiQVxPAmSZJUkLm7VUi73V40rdVq0Wq1ptoO790mSZIAOp0OnU5n7PrIzAk2Z+2IiASY9Ps1lEkC7+UoaWUiAoDMjP55HjaVJEkqyNwdNpWkafBpK5ImxZ43SZKkghjeJEmSCmJ4kyRJKohXm66yg5+2if2PftpEX0Nr0w2XftBtP6emte09P27tWVhYGHgLKjXbNLa7V5tO0Y2f+eCsm6AZcdvPL7f9/FpYWJh1EzQDs97uhjdJkqSCGN4kSZIKUvR93iLiZOCPgIcDtwKfAF6cmZfOsl2SNAneO04SFNzzFhEvAN4DHAfsAG4GWsBFEXH8DJsmSZI0MUX2vEXEQcCrqALbkzPzonr6JuAM4Ezg2Nm1cLK+f/nF3On+j2nE6zTpvUxDkz6vprzGtCznvezpM5ZHvcZq9uZ1Oh1ardaqrW9WrzEtTfq8mvReZqnUnrdTgH2A07vBDSAzzwQuAI6JiKNm1bhJu/nySxrzOk16L9PQpM+rKa8xLbP+vNZv3jrWMI5Op7NaTZ7pa0xLkz6vJr2XWSqy540qvAGcN2DeecBT65o/m1qLJEljBbidl1yFN9iQ9lyp4e1w4PrM/MKAeRfW4yOm2B5J0jLs6aHdce285Co+NOQ1lnMIeKl2XnvN9azfvHUiF4l0X3vUewEvUJlHxYW3qG45fAhw1ZCSnfX40Om0SJJUkkkEx0mH0dV67dUOest930sFURi/jcNee9BrNC3gFvd4rIi4G3Ad8MnM3DBg/j7ALcDnM/NhPdPLeqOSJGnuzcvjsfaqx3vPtBWSJEkTUNxhU6rDoj8EDhoyvzv92t6Jg5KrJElSaYrrecvqOO91wMFDSrrTrx0yX5IkqVjFhbfaFcABQ+7l1n26wpVTbI8kSdJUlBrezqnHJw2Yd2JfjSRJUmMUd7UpQEQcCHy9/vZJAx6PdUlmNuOZOJIkST2K7HnLzF3AZmBfYFtEXBYRV1MFt5uBTdNuU0ScHBHbIuLGiNgZEedHxNHTbocmIyJeFhEfGDE8rK/e/aFQUbkmIl4+omZZ29f9oQxLbfvl/h6ol3Hbr2ER8cyI+EhEXBsR346If42IU4fUrp2f+8wsdgBOBi4CbgJ2Ae8DHj6DdrwAyHr4L+Dq+utbgONn/Tk5rMo2/mzPNh40PM79oRkDsLHeXi8fMn9Z29f9oZxhjG0/9u8Bt/3aH4C31NvjNuBS4FP11wm8l/ro5J5sy0lv+5l/eKUPVLcmuQX4PnBcz/RN9Yb65Kzb6LDibRz19r3U/aG5A7A/8Czgm8P+gC93+7o/lDGMue3H/j3gtl/7A3BcvR2+CjykZ/r9gcvqeb+7J9tyGtu+yMOma8wpwD7A6VmfeweQmWcCFwDHDLkqVuW4J3BH4PIxat0fChQR5wLXA2dRPX5vmOVuX/eHNW4Z2345vwfAbb/WPbse/2X2PCc9My8HuodNf7Mer7mfe8Pbyp1Sj88bMO+8vhqV6f71+Mtj1Lo/lOlC4O+ANwMfGVG33O3r/rD2jbvtl/N7ANz2a90R9fij/TMy89NUp2I9op605n7uS3zCwlpzOHB9b3LvcWE9PmLAPJWj+0v7qxHRprqX4F7AZ4BzM/OrPbXuDwXKzNd2v65PVn7SkNLlbl/3hzVuGdt+Ob8HwG2/1l1Edcj0mv4ZEbEvcGeqJzrBGvy5N7ytQEQEVTf7VUNKuhv+0Om0SBNyZD1+DbBfz/RnAS+OiOdl5lnuD8223O3r/tA4Y/0eALd9CTLzxSNmv4jq+egfW6s/9x42XZmDqQLwriHz/QFthu5/3DuprkY7iOq/phdT3a7m7yPiobg/NN1yt6/7Q7OM+3sA3PZFqm8V8wrgpcCtwCtZoz/39rxN1l71eO+ZtkIrdT7VJeRnZWb35tC7gJdHxO3AK4CXAM9bYj3uD8223O3r/lCWcX8PPGOMdbnt15iIOBZ4A9VVqLcBp2bmZyLibkssOpOfe8PbyuwEfkj1H9gg3enXTqc5moTMfOeI2X9L9Uv7Ubg/NN1yt6/7Q4Ms4/cAuO2LERH7AC8HXkh1NPKzVMHt03XJmvy597DpCmR145brqLpJB+lO9we0oTLzu8C3gfu6PzTbcrev+8P86P09UH/vti9ARNwb2A78MVUv6nOBR/UEtzX7c294W7krgAOG3LPl+Hp85RTbo1UUEXePiFMj4slD5q8D7gJ8pZ7k/tBsy92+7g8NsAe/B8Btv6ZFxP7AVuAo4GLgEZn5psy8fUD5mvu5N7yt3Dn1+KQB807sq1F5bgTeCJwbEXccMP8pVOcuXFp/7/7QbMvdvu4PzbDc3wPgtl/rng88kuoxWE/sOY9xkLX3cz/rR1SUPgAHAjfXw6DHYFw86zY6rHgbv73elu8C7twz/Wiq/7RvAx7m/tCMgeru6sMekbSs7ev+UNawxLYf+/eA237tD8AOqnPT7j5G7Zr7uZ/5B9iEgZ98AO1l7H4A7feBR8+6fQ4r3r4HUj0SJ4HvAB8HPl//4P8AeL77Q3OGUX/A92T7uj+UM4wR3Mf+PeC2X7sD1XlnWYerD4wYzt7TbTnpbT/zD7EpA3Ay1R2bb6I68fF9wMNn3S6HVdu+BwFb6l/WN1H9p/1PwE+7PzRrWCq87cn2dX8oYxgjuC/r94Dbfm0OwLE9wWrUcO1KtuUkt33ULyBJkqQCeMGCJElSQQxvkiRJBTG8SZIkFcTwJkmSVBDDmyRJUkEMb5IkSQUxvEmSJBXE8CZJklQQw5skrUER8daIyCHDjyLimxHxiYj4k4i4yxTbdWrdho9M6zU1nyLiJQP2/V8eUX9CRPxdRHwuInZGxA0RcWlE/FNE/K+I2GcCbXx13a7bIuLAMZf5k3qZ2yPinhHx2gHv84mj1mF4k6TyBHAIcDzwKuBLEfHUgYURH+n+QZhmAyetDpEviYhHzrotmq2IeEhEfBy4AGgDD6V6lNmdgUcCvwL8LfCViHjWKr/8/1+P96Z6HNY4unX/npnX7smLGt4kaW37aGZG7wDsC9wf2AR8HTgUOD8injLLhk7Zs4GXAkfPtBWahrf17P/v650REccBnwAeD9wCvJbqn5qDgAOARwF/BHwDuDfwjoj4tdVqWGZeBny2/nbJ9UbEfYFH19+eU6/jD3p+tneM87qGN0kqTGbemplfycw3AEcBnwLuALw1IvbvK78K+GI9SI0REYcA/xc4ELic6qHvf5iZ2zJzV2bekJmXZubfUP2z83GqXuu3R8QjVrEp3d63J0XEPZao7fa6/QB4756+oOFNkgqWmbuA36m/vQ/wW33zT83Mh2TmQ6beOGmy/oaq1/m7wJMz8/JhhZl5E/CLwHVU/+i0V7Ed5wBJlal+dYnap9fjD2Xmzj19QcObJBUuMz8NfKj+9ifCW89J328dtGxEPKA+yfsrEXFzRHwtIj4QEb8UEbHctkTEvSLibyLiCxHx/Yj4RkR8NCKeFRHrRix3SET8dUR8sV7umvp8vWf1tqN7IQfws/Wkt/W/v4i4sp52cETcPyI6EXFj/2cQEQdFxEvr9/v1iLipbvfrIuI+Q9rZvZDkGVH5vYj4ZH1y/Dci4t8iYuMefG69bX5wRPxjfVHKzRFxWUT8z57an68/m+/Ww7aIOGnEuo+KiH+IiP+KiOvr4XMR8faI2DBiuSPrfeMzPe9vW0RsiiEn/0fEPSLiNRGxPSJ2RcS3I+KSiHjxgF7hPVZvn+5hyj/PzK8ttUxmfg94I/Bt4OeG7Y8R8fSIeH9EfKveFz8fEX8bEfcfst6vAf9ef3vKiDbfG3hs/e05S7V3pMx0cHBwcFhjA/BWqv/mPzJm/a/X9bcDB/RMf0k9/a0DlvkN4LZ6/qDh3cAd+pY5dVi7gJ8Dvjdiff8O3GXAco8Bdo5Y7oPAT/V9Lv3DW3vWd2U97XjgW0NqjqE6D2rYa94IPG3EdjmF3T0ug4ZNy9ze3Tb/4ojP8A+APxzxmr8yYL2/CPxoxDK3D1luI9WhvWHL/Qewrm+ZY+rPbdgyXwb2X8ZnMmrf/f2e7XSnVfqZ2xc4b0T7bwWeMWTZ3+6pWz+kZlM9//vDPoee/eCJo9pqz5skNcOOevxTwMOXKo6Ix1EFkb2BjwFPBO5CdQ7da6j+4P8q8PpxXrzulXgvsD/VCdw/DxwMPBD4U6o/fE8A3tW33L5UIfEgqj/uG+vl7gNspgoQTwVeBJCZv5XVid0frVfx7KxO9v6JHsfa2VTh5NlUVyD+Xv2adwXeR3XIbRdVb+X9gHtQXZl4FbAfcG5EHD7kLb8AeAbV1b4Pozo5/inAF+r5r46IOwz7vEY4q17H46g+y8f1rHML1ba5ADi2fs2fB7pXLL6od0X1659FdZ7Xx6nC7AFU2/kJwEVU+8vr+pa7M9Vntw74MFW43o/qCufnUn2mjwOe09f2c+u6TwFPql/rIKrP6Qaq887+fHkfx1BPqMefyMzvr9I6XwecCPwQeAXVPnNXqu16MdXh1rMj4rEDln0P1T4O1fsdpHu+2z9n5g0raulqpFUHBwcHh9UdWH7P233Z/Z//L/VMH9h7AWyrp/8Lda9W3/zfZXfPzEN7pg/seWN3L9TngH0HrO+pPe17as/0P66nXQscOGC5P6/nf6lv+kfq6acOWKbbe3EDcMiA+afV828DHjxg/t2ozo1K4M1DtksCLxiw7EN75h+zjO3dbfMVwB375j2hZ53b+rcXVThN4Ht904+up98C3HXAa96vZ71375n+ZHb3NO03YLl31vPf3jPtiJ51PWDAMq+o5/37Mj6TUT1vF9fzzlyln7cHs7uH8jcGzF9H9U9OAtuGrOPcev6lA+YdWv8sJXDSGPvBE0e11543SWqGb/Z8PfLcoog4DDiu/nZTZv5oQNkC8F9UPTMjz+GKiL2A7jlXL8zMW/prMvMCYGv97S/1zPq9enxGVhdf9Hsr1ZWEPxURB41qxwBvycxvDZjefT9vy8xFV+Fm5rfZ3Rv1i0PW/VWq86f6l/08VfCDJbbDEGdk5s190/6z5+tXD9heFw95veuozgs7MTO/O+C1eu8xdqchX99xwHL/CzgMeOGQZfYdsMwr62WG9UotV/eGuEN7sCLir2L4ja4zIk7tKX86VQ/lxZl5Vv+6MvOHVIesAY6L6krXfu+sx4+MiP4LhE6i+ln6HvCBpd/eaIY3SWqG3j8mSx2SeVA9/vqg8AKQVTfAJ+pvf3qJ9a0H9qE63PSxEXX/0bu++rDe+nrahwYtkJnXZOYD6uE7S7Sj36VDpndPPL9gxLL/Uo/vFRGDAsy2+g/6IDeN0bZhPt8/oS/MfW7AMgMPG9af3T9m5gd7p0fEXlHdH+3MIW24hCpk3AHYHhGnRXV/su56v1ev+9s9y/w30L1o4MMR8cqIeHDPMjfVy3xjyGsuV/czHnhhyR7o/kz864iaT7H7sx50ocf7qQ7Dw+J7vnWvMj1v0D83y2V4k6Rm6P0jttQtCB5Yj+81qmeC6tAp7O7lWGp964CbRqzvr/rWdyS7/w5dtcRr7IlFQSEiDmB30N0xYtnuvAAGnfd2xUoaNsKtK5y/SEQcX4epTkR8nir4bGPI7TLq3spfo/r87kd1yPOr9RWx76ivxrxT3zK3UZ3TdTnV5/unwBfqq3jfHRG/WZ9ruFq6PaoPGlaQmX+afTe4zup8yW0Dyrv78Gkj9t/b2d3DuOhnov4M3l1/++OrTiPibuy+OvpdrALDmyQ1Qzdg/Ijdd3wfZrnPeFzq8N+erq/38Npty1zHOJYKOjliXm+v2t4D5k+ivasqIu4UER2qHtQ/pTrv8Gbg7VTB7dhhy2bmB6gCzW9TXdxxA1Uv6W9Qndv13xHxc33L/CfVBS/dK3F3AvekuvDl7cAVEfHMVXp7F9XjB0fEoO0zUEQE1XmJ/VbrZ6J7w94HRMQx9dcnAntR3aLkw8t8nYEMb5LUDL9Zjz+T1f2sRvnvevylQT0TA4bjRq5t9/puBvYaY33du9D33lT1fuO/1T1Xfzbdc9LWjyg9oufroTd/XeO2UJ2z9y2qAHXnzDw2M383M9+SmZ8atXBWTyj4h8w8kaqn6VjgZcDVVI+aenfdq9S7zK2Z+a7MfCZwd6orcV9EtY8cSHVfvkHhabneX4/3Z+kb4/Z6MNXVtv26+/DA3roBw5uHrP8T7O617R467V5leu6IQ+3LYniTpMJFxKOAE+pv/2GMRbp/qO4fEfuNWO8DIuLYiLjnEuv7ClWP3x2BB4xY333r9d0PqnDA7kObTxiyzD5R3az2hoh49KCaPfCVenzCiJpur9I1Wd2dv0T/ox6flpnvqQ/r/diwm85GxMaIeGFE/EJ3WmbenpmfysyXUD3s/Taq22gcVy/z+HqZU3qWycz8fGa+GngEcA1VL+aTV/rGMvNidp+T+epxLmape92G3fqm+zPxyBHLr6v332OH9fbV54qeXX/7a3W7us8cXpVDpmB4k6SiRcSBwN/X314FvG2Mxb4CfInqUM5Lh6z3PsCnge2MCGRQ9baw+3DQ6UPWtz9wYb2+J/bM6v6h+5P6fLR+z6A6hyrq9ixa9ai2DdHttfmtiHhg/8z6SsLulZQrvjJwhrqHAoedIN97X7jez/HxVPeTOyMiBuWEH/R8/d16/LB6mbcM+YfgBz2v8d0B8/fE86kOjd+b6iKJew0rjOqJEK9hdyjv170S+uT6n6FBNlHtv+/mJw+r9+seOr0P8GqqwHo11b32VoXhTZIKExF7R8ThEfFcqvPbHkX1R+y3MvPGpZavewc219++MKpHPj0qIvaL6pFRJ1NdNbofcFFmjrqCtOvPqf6gPT0izo+Ix0bEARFxl4h4ar2+e1MdUnp3z3KnA9+h+kN3cUQ8LSIOjIi7R8TvAG+o696Ymb2hoeuYukdlOd4AfJ3qasptEfEbEXFYRBwaEb9C9Qf6YKrzvF65zHWvJd3Doi+NiCfU58AdHBFPiYj3s/uZuACt2P3Iq+4J/UcCZ0XEQ+oe0P0j4meAf6b67L5D9VlBdQ5aUu0z50XEMRFxx3qf2kB1E9t7Ue2n3UdJrUh92Pd3qPa7RwFfiogtEXFcvQ/tV7fjf1Ld9uYPqa5AXhSiMnMb1bl9ewMfiojnRsT6+n0fEREvA/66Lv/r+mdoWLu+wO7P/rfr8T+OWmbZRt0EzsHBwcFhNgPDHwM1aPgmPTe+7VvPqBud/gmjH530WeBufcsMvElvPe/XGf24rauBIwcs9/NUV0AOW+7DwN59y/yfnvk3AP+nZ96SNzqlutXDt0a85g2DPtOe7fKSEese60ary1mmp133GzDvxzfb7Zt+LFWv26D39yOqYPpvPdM+3bPsu5bY524Gfqbv9bYsscztDHm81HL33b66E6j+KRj12j+ietrE/lQXbyR9N3imOgz8H0usZ8uYbf+DvuU2rMZ+0B2GPiRYkrRmJdWVfJdT9Ra8OQffhHX0SjJfHRH/RvWYo0dSHfq6vl7vWcA7cnBv17D1vTMi/hN4HlVPyMOpelq+QvXorDflgEcZZeYHIuJhVIfBNlCdH3UD8EWqw8BnZ/2XrcdLqG76+niqQ3LXsQyZub0+cX4T1eOfHkHVm3QZ8Emqm+Uu+bDztSwzPxkRxwJ/ATyaqufrOqqnU7y+nn8E1WG+B9bTu55Jtc3aVBdv3IvqcOdXqcL0mdl3z7bM3BwRH6H6TB9ItX1uojqc/4n6Nb/CKsvMD0XEg6j+efhlqidLHEoVXL9et/cdmbkdICIuGbKe70bEz1I9seKpVPvEYXX7LwX+pruOMbyLqqduL+DyZSw3llj88yBJkjRbEfESqnMy35aDn13bOBFxJdVV0D+bmUMPL3vOmyRJUkEMb5IkSQUxvEmSJBXE8CZJktayZ/c8Y/SXZ92Y1RYRr+15fur6cZYxvEmSJBXEq00lSZIKYs+bJElSQQxvkiRJBTG8SZIkFcTwJkmSVBDDmyRJUkEMb5IkSQX5f3VnP9CdMTGIAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "ax.set_xlabel(r'Dielectron mass [GeV]')\n",
    "\n",
    "pairs = ak.combinations(dyevents.Electron, 2, fields=[\"lep1\", \"lep2\"])\n",
    "mass = (pairs.lep1 + pairs.lep2).mass\n",
    "ax.hist(ak.flatten(mass), bins=np.linspace(0, 200, 50));"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Ground truth\n",
    "If we want to identify variables useful to distinguish true electrons and photons from background, we better have a good definition of \"true\".  Let's take a look at the generated particles that are matched to reconstructed electrons and photons to understand what kind of processes give rise to them.\n",
    "\n",
    "For electrons and photons, NanoAOD pre-computes the truth-matched particle by finding final-state electron and photon particles in the `GenPart` table that are spatially close (via the metric $\\Delta R = \\sqrt{(\\Delta \\eta)^2 + (\\Delta \\phi)^2}$) to the reconstructed particle.  When no match is found, we get a null/NA value.  Below, we put reconstructed electrons into three buckets: matched to a gen electron, matched to a gen photon, and unmatched, based on the [PDG id](http://pdg.lbl.gov/2018/reviews/rpp2018-rev-monte-carlo-numbering.pdf).  Since for this purpose we don't care so much which event each electron belongs to, we also flatten the array here for convenience."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "electrons = ak.flatten(dyevents.Electron)\n",
    "\n",
    "true_electrons = electrons[ak.fill_none(abs(electrons.matched_gen.pdgId) == 11, False)]\n",
    "photon_fakes = electrons[ak.fill_none(electrons.matched_gen.pdgId == 22, False)]\n",
    "unmatched_electrons = electrons[ak.is_none(electrons.matched_gen)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For the true electrons, we would expect the parent to be a Z boson in these samples.  We can check this using the `distinctParent` attribute stored in our files.  There is also a `parent` attribute, which specifies the direct ancestor to the generated electron, however this could also be an electron (due to final state $\\gamma$ radiation and other generator-internal reasons).  The `distinctParent` attribute skips copies of the same particle that are the result of FSR.\n",
    "\n",
    "Below, we put the `distinctParent` value, along with some other (carefully chosen) gen particle flags, into a pandas dataframe by passing a dictionary. In this dictionary, the key signifies the parameter and the element is a flattened array of that value for every electron in the NanoAOD file. The `ak.zip` groups keyed elements together for each electron, making sure the dataframe is cast correctly where each electron is an entry. We then use `groupby` to create categories that share common values. They are then ranked by the number of electrons, or \"counts\" that are in that category. We only consider the categories with the top 10 number of entries."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>parentId</th>\n",
       "      <th>fromHardProcess</th>\n",
       "      <th>isPrompt</th>\n",
       "      <th>isDirectPromptTauDecayProduct</th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>23.0</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>39213</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>15.0</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>5031</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2.0</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>1130</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>329</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>4.0</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>258</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>21.0</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>196</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3.0</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>122</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>111.0</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>511.0</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>521.0</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    parentId  fromHardProcess  isPrompt  isDirectPromptTauDecayProduct  count\n",
       "15      23.0             True      True                          False  39213\n",
       "11      15.0            False     False                           True   5031\n",
       "3        2.0             True      True                          False   1130\n",
       "1        1.0             True      True                          False    329\n",
       "6        4.0             True      True                          False    258\n",
       "13      21.0             True      True                          False    196\n",
       "4        3.0             True      True                          False    122\n",
       "16     111.0            False     False                          False     58\n",
       "22     511.0            False     False                          False     50\n",
       "23     521.0            False     False                          False     47"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = ak.to_pandas(\n",
    "    ak.zip(\n",
    "        {\n",
    "            \"parentId\": abs(true_electrons.matched_gen.distinctParent.pdgId),\n",
    "            \"fromHardProcess\": true_electrons.matched_gen.hasFlags([\"fromHardProcess\"]),\n",
    "            \"isPrompt\": true_electrons.matched_gen.hasFlags([\"isPrompt\"]),\n",
    "            \"isDirectPromptTauDecayProduct\": true_electrons.matched_gen.hasFlags(\n",
    "                [\"isDirectPromptTauDecayProduct\"]\n",
    "            ),\n",
    "        }\n",
    "    )\n",
    ")\n",
    "\n",
    "top10 = (\n",
    "    df.groupby(list(df.columns))\n",
    "    .size()\n",
    "    .rename(\"count\")\n",
    "    .reset_index()\n",
    "    .sort_values(\"count\", ascending=False)\n",
    ")[:10]\n",
    "\n",
    "top10"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Indeed, in most cases, the parent is pdgId `23`, corresponding to the Z boson.  Sometimes the parent is a $\\tau$ lepton, and then the remainder are cases where the electron is from a hadron decay (`parentId > 100`) or where the parentage appears somewhat suspect (e.g. pdg IDs less than 6, corresponding to bare quarks... this is a quirk in how MadGraph records the parentage history for off-shell $Z^*/\\gamma^*$)  Luckily, a set of flags is embedded in NanoAOD that are more straightforward than travsersing the NanoAOD parentage history.  In our case, we'll take `isPrompt` as the marker for true electrons coming from the hard scatter event.  Since direct $\\tau$ decays produce prompt isolated electrons, we include those with `isDirectPromptTauDecayProduct` as well."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "promptmatch = true_electrons.matched_gen.hasFlags(['isPrompt'])\n",
    "promptmatch = promptmatch | true_electrons.matched_gen.hasFlags(['isDirectPromptTauDecayProduct'])\n",
    "prompt_electrons = true_electrons[promptmatch]\n",
    "nonprompt_electrons = true_electrons[~promptmatch]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There is also a pre-computed variable `Electron.genPartFlav`, which according to the documentation, has the assignment:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Flavour of genParticle (DressedLeptons for electrons) for MC matching to status==1 electrons or photons: 1 = prompt electron (including gamma*->mu mu), 15 = electron from prompt tau, 22 = prompt photon (likely conversion), 5 = electron from b, 4 = electron from c, 3 = electron from light or unknown, 0 = unmatched'"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dyevents.Electron.genPartFlav.__doc__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's see how this compares to our analysis: below we count the number of occurrences of each flavor using [np.unique](https://docs.scipy.org/doc/numpy/reference/generated/numpy.unique.html) and then print the results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "GenPart flavor   1 has    41307 occurrences\n",
      "GenPart flavor  15 has     5031 occurrences\n"
     ]
    }
   ],
   "source": [
    "flavors, counts = np.unique(ak.to_numpy(prompt_electrons.genPartFlav), return_counts=True)\n",
    "for flavor, count in zip(flavors, counts):\n",
    "    print(\"GenPart flavor % 3d has % 8d occurrences\" % (flavor, count))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Try checking our other categories against this pre-computed ground truth categorization by switching out the `prompt_electrons` variable with the others.\n",
    "For photons, `gevents.Photon.genPartFlav` provides similar categorization.  Below, we use this variable to sort photons into two variables."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "photons = ak.flatten(gevents.Photon)\n",
    "prompt_photons = photons[photons.genPartFlav == 1]\n",
    "unmatched_photons = photons[photons.genPartFlav == 0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Discriminating variables\n",
    "In the following few cells, we'll introduce typical discriminating variables used to distinguish true hard-process electrons and photon from fakes, and plot them.  Later, we'll look at electron- and photon-specific discriminating variables."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Shower shape\n",
    "\n",
    "Shower shape variables are constructed from the energy-weighted spatial distribution of ECAL and HCAL deposits.\n",
    "\n",
    "   * $\\sigma_{i\\eta i\\eta}$ (`Electron.sieie`) is the energy-weighted standard deviation of integer eta (one per crystal) within the 5×5 crystal cluster centered at the crystal with maximum energy:\n",
    "\n",
    "$$ \\sigma_{i\\eta i\\eta} = \\frac{\\sum_{i}^{5x5} w_i (i\\eta_i - i\\eta_{\\text{seed}})}{\\sum_i^{5x5} w_i}, \\quad w_i = \\max(0, 4.7 + \\ln E_i - \\ln E_{5x5}),$$\n",
    "\n",
    "where the sum runs over the 5 × 5 crystal matrix around the most energetic crystal in the SuperCluster, and the eta distances are measured in units of the crystal size in the eta direction. This variable represents the second moment of the energy distribution along the eta coordinate.  The integer values are scaled by the nominal crystal $\\phi$ width in the barrel which is $\\pi / 180 \\approx 0.0175$.  PS it is defined [here](https://github.com/cms-sw/cmssw/blob/master/RecoEcal/EgammaCoreTools/interface/EcalClusterLazyTools.h#L252) and implemented [here](https://github.com/cms-sw/cmssw/blob/master/RecoEcal/EgammaCoreTools/interface/EcalClusterTools.h#L1034)\n",
    "\n",
    "   * $R_9$ (`Electron.r9`) is the ratio of the energy deposited in ECAL inside a 3 × 3 crystal matrix centered on the most energetic crystal in the supercluster and the superCluster raw energy.\n",
    "\n",
    "   * $H/E$ (`Electron.hoe`) is the ratio of the energy deposited in the single closest HCAL tower to the supercluster position inside a cone of radius 0.15 in the $\\eta,\\phi$ plane centered on the photon direction and the energy deposited in ECAL to that supercluster.\n",
    "\n",
    "Below, we plot the $\\sigma_{i\\eta i\\eta}$ variable for true prompt electrons and unmatched electrons, where the minimum electron $p_{T}$ is required to be at least 20 GeV."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "ax.set_xlabel(r'Electron $\\sigma_{i \\eta i \\eta}$')\n",
    "common_opts = {\n",
    "    'alpha': 0.5,\n",
    "    'bins': np.linspace(0, 0.05, 100),\n",
    "    'density': True,\n",
    "}\n",
    "\n",
    "def cut(candidates):\n",
    "    return candidates[\n",
    "        (candidates.pt > 20.)\n",
    "    ]\n",
    "\n",
    "ax.hist(cut(prompt_electrons).sieie, color='blue', label='Prompt electrons', **common_opts)\n",
    "#ax.hist(cut(photon_fakes).sieie, color='green', label='Photons', **common_opts)\n",
    "#ax.hist(cut(nonprompt_electrons).sieie, color='orange', label='Non-prompt electrons', **common_opts)\n",
    "ax.hist(cut(unmatched_electrons).sieie, color='red', label='Unmatched', **common_opts)\n",
    "\n",
    "ax.legend()\n",
    "ax.set_ylabel('Normalized' if common_opts['density'] else 'Events');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Try uncommenting different categories and looking at the distributions.  Any guess as to why the true electrons have a double-peak?  It might help to try looking at `sieie` separately for barrel and endcap reconstructed electrons (hint: modify the filter function to also require `abs(electrons.eta) < 1.4447`).\n",
    "\n",
    "**Duplicate the above cell and look at R9 and H/E variables.**  P.S. both variables have a natural range of 0 to 1.\n",
    "Which of the three is your favorite?\n",
    "\n",
    "Look at each variable as a function of $p_{T}$.  Simple option: draw two histograms, one with electron $p_{T} < 20$GeV cut and one with electron $p_{T} > 80$GeV); fancy option: 2D histogram.  What trends do you notice?\n",
    "\n",
    "Plot `sieie` for prompt and unmatched photons in the barrel.  Notice the double peak for unmatched?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Isolation\n",
    "\n",
    "Isolation variables compare the amount of energy deposited in a region around the reconstructed object to that of the object.  Typically, non-prompt electrons and photons (i.e. from hadron decays) have other energy deposited nearby.  Two particle flow (PF) isolation variables are saved in NanoAOD:\n",
    "\n",
    "   * PF charged hadron isolation `pfRelIso03_chg`: sum of $p_{T}$ of PF charged hadrons which are associated to the primary vertex, not in the \"footprint\" of the candidate, and within a cone of $\\Delta R < 0.3$,\n",
    "   * PF total isolation `pfRelIso03_all`: sum of $p_{T}$ of all PF candidates within the cone of $\\Delta R < 0.3$,\n",
    "   \n",
    "where in both cases, the values are relative to the object $p_{T}$.\n",
    "\n",
    "Due to pileup (multiple collisions in the same event), some of the PF candidates within this isolation cone may belong to a different collision and therefore this variable has a mild dependence on the number of pileup interactions.  A correction is made to the total isolation to remove this pileup-dependence,\n",
    "\n",
    "$$I = I_{\\text{raw}} - \\rho * \\text{EA}(|\\eta|),$$\n",
    "\n",
    "where $\\rho$ is the median pileup energy density per unit area, a measure of the pileup activity in the event (computed during the jet clustering stage with FastJet, acting on the PF candidates); and EA is the effective area of the electron or photon object, a quantity computed and [made available](https://github.com/lsoffi/cmssw/blob/CMSSW_9_2_X_TnP/RecoEgamma/ElectronIdentification/data/Fall17/effAreaElectrons_cone03_pfNeuHadronsAndPhotons_92X.txt) by the Egamma POG, and is defined as explained in [slide 4 here](https://indico.cern.ch/event/662749/contributions/2763091/attachments/1545124/2424854/talk_electron_ID_fall17.pdf)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "ax.set_xlabel(r'Electron relative charged isolation')\n",
    "common_opts = {\n",
    "    'alpha': 0.5,\n",
    "    'bins': np.linspace(0, 1, 100),\n",
    "    'density': True,\n",
    "}\n",
    "\n",
    "ax.hist(prompt_electrons.pfRelIso03_chg, color='blue', label='Prompt electrons', **common_opts)\n",
    "# ax.hist(photon_fakes.pfRelIso03_chg, color='green', label='Photons', **common_opts)\n",
    "# ax.hist(nonprompt_electrons.pfRelIso03_chg, color='orange', label='Non-prompt electrons', **common_opts)\n",
    "ax.hist(unmatched_electrons.pfRelIso03_chg, color='red', label='Unmatched', **common_opts)\n",
    "\n",
    "ax.legend()\n",
    "ax.set_yscale('log')\n",
    "ax.set_ylabel('Normalized' if common_opts['density'] else 'Events');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Electron-specific variables\n",
    "\n",
    "   * Mini-isolation `miniPFRelIso_all,miniPFRelIso_chg`: uses a cone whose size varies as a function of the electron $p_{T}$ in order to reduce accidental overlap from hard scatter and pileup processes,\n",
    "   * Conversion veto `convVeto`: reject photons that converted in the innermost layers of the pixel or tracker detectors"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Photon-specific variables\n",
    "\n",
    "   * Conversion-safe electron veto `electronVeto`: this cut replaces the detector-based `pixelSeed` used to distinguish between photons and electrons. It checks the $\\Delta R$ distance between the photon direction and the nearest electron track and remove photons if there is a GsfElectron matching to the photon supercluster, with no missing hits and having no matching reconstructed conversion."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Cut optimization\n",
    "\n",
    "Now that we've seen a few discriminating variables, let's compare their power for differentiating true electrons from background using a [ROC curve](https://en.wikipedia.org/wiki/Receiver_operating_characteristic), which shows signal tagging efficiency as well as background rejection efficiency simultaneously on a 2D plot. Efficiency $\\epsilon$ is defined for a given selection as:\n",
    "\n",
    "$$\\epsilon = \\frac{n_{\\text{pass}}}{n_{\\text{pass}}+n_{\\text{fail}}} = \\frac{n_{\\text{pass}}}{n_{\\text{total}}}$$\n",
    "\n",
    "Below we plot the ROC curve for: selected discriminating variables; working points in the ROC of the POG-supported cut-based ID, which is a tuned selection on multiple variables; and the ROC for the POG-supported MVA classifier. Information on all of the POG-supported IDs can be found [here](https://twiki.cern.ch/twiki/bin/view/CMS/EgammaIDRecipesRun2).\n",
    "\n",
    "**Define a selection on multiple variables, measure its signal and background efficiency, and place a marker on the ROC corresponding to your ID**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "def makeroc(sig, bkg, binning, direction=1):\n",
    "    '''Given a signal and background discriminant array, along with a binning,\n",
    "    compute the ROC curve by integrating discriminant histograms along a direction:\n",
    "        1: cut < value\n",
    "        -1: cut >= value\n",
    "    Returns a tuple of (signal efficiency, background efficiency)'''\n",
    "    sighist, _ = np.histogram(ak.to_numpy(sig), bins=binning)\n",
    "    bkghist, _ = np.histogram(ak.to_numpy(bkg), bins=binning)\n",
    "    # numpy histogram will ignore values out of range of our binning (overflow)\n",
    "    # so add the overflow values that would pass given our cut direction\n",
    "    if direction == 1:\n",
    "        sighist[0] += sum(sig < binning[0])\n",
    "        bkghist[0] += sum(bkg < binning[0])\n",
    "    elif direction == -1:\n",
    "        sighist[-1] += sum(sig > binning[-1])\n",
    "        bkghist[-1] += sum(bkg > binning[-1])\n",
    "    else:\n",
    "        raise ValueError(\"Invalid direction\")\n",
    "    sigpass = (np.cumsum(sighist[::direction]) / len(sig))[::direction]\n",
    "    bkgpass = (np.cumsum(bkghist[::direction]) / len(bkg))[::direction]\n",
    "    return (sigpass, bkgpass)\n",
    "\n",
    "\n",
    "def cut(electrons):\n",
    "    return electrons[\n",
    "        (electrons.pt > 20)\n",
    "        & (abs(electrons.eta) < 1.4)\n",
    "    ]\n",
    "\n",
    "\n",
    "roc_sieie = makeroc(\n",
    "    sig=cut(prompt_electrons).sieie,\n",
    "    bkg=cut(unmatched_electrons).sieie,\n",
    "    binning=np.linspace(0.005, 0.02, 50),\n",
    "    direction=1,\n",
    ")\n",
    "\n",
    "roc_hoe = makeroc(\n",
    "    sig=cut(prompt_electrons).hoe,\n",
    "    bkg=cut(unmatched_electrons).hoe,\n",
    "    binning=np.linspace(0., 0.3, 50),\n",
    "    direction=1,\n",
    ")\n",
    "\n",
    "roc_mva = makeroc(\n",
    "    sig=cut(prompt_electrons).mvaFall17V2Iso,\n",
    "    bkg=cut(unmatched_electrons).mvaFall17V2Iso,\n",
    "    binning=np.linspace(-1, 1., 50),\n",
    "    direction=-1,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "sig_pass = cut(prompt_electrons).cutBased >= prompt_electrons.LOOSE\n",
    "bkg_pass = cut(unmatched_electrons).cutBased >= prompt_electrons.LOOSE\n",
    "wp_cutbased_loose = (\n",
    "    sum(sig_pass) / len(sig_pass),\n",
    "    sum(bkg_pass) / len(bkg_pass),\n",
    ")\n",
    "\n",
    "sig_pass = cut(prompt_electrons).cutBased >= prompt_electrons.TIGHT\n",
    "bkg_pass = cut(unmatched_electrons).cutBased >= prompt_electrons.TIGHT\n",
    "wp_cutbased_tight = (\n",
    "    sum(sig_pass) / len(sig_pass),\n",
    "    sum(bkg_pass) / len(bkg_pass),\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f2c738df100>"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "ax.plot(*roc_sieie, label=r\"Cuts on $\\sigma_{i\\eta i\\eta}$\")\n",
    "ax.plot(*roc_hoe, label=r\"Cuts on $H/E$\")\n",
    "ax.plot(*wp_cutbased_loose, label=r\"POG loose ID\", marker='*', markersize=20, linestyle='none')\n",
    "ax.plot(*wp_cutbased_tight, label=r\"POG tight ID\", marker='*', markersize=20, linestyle='none')\n",
    "ax.plot(*roc_mva, label=r\"POG MVA ID\")\n",
    "\n",
    "diag = np.linspace(0, 1, 100)\n",
    "ax.plot(diag, diag, color='grey', linestyle='--')\n",
    "ax.set_xlabel(\"Signal efficiency\")\n",
    "ax.set_ylabel(\"Background efficiency\")\n",
    "ax.set_xlim(0, 1)\n",
    "ax.set_ylim(1e-3, 1)\n",
    "# ax.set_yscale('log')\n",
    "ax.legend(title=r'Barrel electrons')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We are also often interested in the efficiency as a function of the electron's kinematics. To help explore the dependency, we can record the number of electrons passing and failing a given selection separately for many small regions in the electron $p_{T}$ and $\\eta$, which will allow us to compute the (binned) efficiency as a function of these variables, $\\epsilon(p_{T}, \\eta)$. We are using the [hist](https://hist.readthedocs.io/en/latest/) package to make a multidimensional histogram to record all these values in one object, which later can be projected as necessary."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Hist(\n",
       "  StrCategory(['pass', 'fail'], name='region', label='Region'),\n",
       "  Regular(18, 20, 200, name='pt', label='$p_{T}$'),\n",
       "  Regular(25, -2.5, 2.5, name='eta', label='$\\\\eta$'),\n",
       "  storage=Double()) # Sum: 36483.0 (46338.0 with flow)"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import hist\n",
    "\n",
    "efficiency = hist.Hist(\n",
    "    hist.axis.StrCategory([\"pass\", \"fail\"], name=\"region\", label=\"Region\"),\n",
    "    hist.axis.Regular(18, 20, 200, name=\"pt\", label=r\"$p_{T}$\"),\n",
    "    hist.axis.Regular(25, -2.5, 2.5, name=\"eta\", label=r\"$\\eta$\")\n",
    ")\n",
    "\n",
    "passing = prompt_electrons.cutBased >= prompt_electrons.TIGHT\n",
    "efficiency.fill(region=\"pass\", pt=prompt_electrons[passing].pt, eta=prompt_electrons[passing].eta)\n",
    "efficiency.fill(region=\"fail\", pt=prompt_electrons[~passing].pt, eta=prompt_electrons[~passing].eta)\n",
    "\n",
    "efficiency"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the cell below, we project the histogram to recover the binned $n_{\\text{pass}}(p_{T})$ (as `num`) and the binned $n_{\\text{total}}(p_{T})$ (as `denom`) and plot the ratio $\\epsilon(p_{T})$. Modify which axis the histogram is projected onto to see the efficiency as a function of $\\eta$.\n",
    "**Q: is the efficiency symmetric in $\\eta$?**\n",
    "**Q: what might explain the lower efficiency near $|\\eta|=1.4$?**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 1, '$\\\\epsilon$')"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from hist.intervals import clopper_pearson_interval\n",
    "\n",
    "num = efficiency[\"pass\", ...].project(\"pt\").view()\n",
    "denom = efficiency.project(\"pt\").view()\n",
    "edges = efficiency.axes[\"pt\"].edges\n",
    "err = abs(clopper_pearson_interval(num, denom) - num / denom)\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "mplhep.histplot(num / denom, bins=edges, yerr=err, ax=ax)\n",
    "ax.set_xlabel(\"$p_{T}$ [GeV]\")\n",
    "ax.set_ylabel(\"$\\epsilon$\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Tag and probe\n",
    "\n",
    "So far, we've computed all our efficiencies using ground truth as provided by our simulation to categorize our signal and background reconstructed electrons.  If we trust our simulation to perfectly describe the detector response, then our job is done.  We cannot do that (**Q: why might we need to know true electron tagging efficiency in data?**), so we need to measure the efficiency in data, categorizing signal and background by inference.  How do we do this without selection bias?  Thanks to the Z boson, we have a solution: tag one electron in an event that passes some high quality requirement, then look for a \"probe\" electron such that the invariant mass of the di-electron pair ($m_{ee}$) matches that of the Z boson.  If we count the number of probes falling in some mass window that pass and fail our ID, we can infer the true electron tagging efficiency in data, despite not knowing the ground truth for each electron.\n",
    "Below is a basic example of making a tag and probe selection, then plotting the pass and fail mass distributions, for POG cut-based electron ID, using our simulated Z event sample."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "zcands = ak.combinations(dyevents.Electron, 2, fields=[\"tag\", \"probe\"])\n",
    "# filter pairs to have a good tag\n",
    "zcands = zcands[\n",
    "    (zcands.tag.pt >= 20)\n",
    "    & (abs(zcands.tag.eta) < 2.1)\n",
    "    & (zcands.tag.cutBased >= zcands.tag.TIGHT)\n",
    "    & (zcands.probe.pt >= 20)\n",
    "]\n",
    "# filter out events that have no z candidates\n",
    "zcands = zcands[ak.num(zcands) > 0]\n",
    "# some events may have multiple candidates, take the leading one (as they are sorted by pt)\n",
    "zcands = ak.firsts(zcands)\n",
    "# compute invariant mass\n",
    "mass = (zcands.tag + zcands.probe).mass\n",
    "\n",
    "goodprobe = zcands.probe.cutBased >= zcands.probe.LOOSE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(1, 0, 'Dielectron mass [GeV]')"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (axp, axf) = plt.subplots(1, 2, sharey=True, figsize=(12, 6))\n",
    "mbins = np.linspace(60, 120, 61)\n",
    "\n",
    "axp.hist(mass[goodprobe], bins=mbins)\n",
    "axp.set_title('Passing probes')\n",
    "axp.set_ylabel('Events')\n",
    "axp.set_xlabel('Dielectron mass [GeV]')\n",
    "\n",
    "axf.hist(mass[~goodprobe], bins=mbins)\n",
    "axf.set_title('Failing probes')\n",
    "axf.set_xlabel('Dielectron mass [GeV]')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now make the same plot with low and high probe $p_{T}$ requirements.  How does the mass shape and relative peak height change?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Plot the same distribution for electrons in real data**\n",
    "\n",
    "Use your DAS skills to find a file from `/EGamma/Run2018A-UL2018_MiniAODv2_NanoAODv9-v1/NANOAOD` and open it, then make the same selection and plot it.\n",
    "\n",
    "You'll notice the failing distribution's peak is hidden under a much larger amount of background. This is combinatorial background from random fake electrons combining with the tag to create a continuous falling invariant mass spectrum.  To properly measure the peak height, a full analysis would have to subtract this background by fitting the pass and fail distributions.  If you got this far and are bored, try to use [scipy.optimize.curve_fit](https://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.curve_fit.html) to fit each distribution.  Bonus points if you fit them using a simultaneous log likelihood."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Scale Factors"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Tag and probe is the main method used derive the efficiencies that serve as inputs to the electron and photon *scale factors* (and most scale factors in general). The scale factors we will be using are defined as the ratio of the data efficieny to the Monte Carlo (MC) efficiency.\n",
    "\n",
    "\n",
    "Scale factors for electrons and photons are applied as event weights. Scale factors serve to match the MC object efficiency (ID, reconstruction, etc.) to the observed data efficiency. **Q: Why could this correction be needed?**. Scale factors are accessible via [ROOT Histograms](https://twiki.cern.ch/twiki/bin/view/CMS/EgammaUL2016To2018#General_note_about_ID_SFs) and [JSONs](https://twiki.cern.ch/twiki/bin/viewauth/CMS/EgammaSFJSON). The JSON format works particularly well with NanaAOD and Coffea-based analyeses, so we will include a brief example of how to access the electron ID scale factors using the `correctionlib` pacakge."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "from correctionlib import CorrectionSet\n",
    "\n",
    "evaluator = CorrectionSet.from_file(\"/cvmfs/cms.cern.ch/rsync/cms-nanoAOD/jsonpog-integration/POG/EGM/2018_UL/electron.json.gz\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that we have scale factors accessible, we need to apply them to our analysis! We already have a collection of Z candidates built (our tag and probe pairs), so we will use those. Normally you will not apply SFs to a tag and probe analyses, but this is just to serve as an example of how to apply scale factors to a Z candidate. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(1, 0, 'Dielectron mass [GeV]')"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Get the ID scale factor for each electron in the pair\n",
    "sfel1 = evaluator[\"UL-Electron-ID-SF\"].evaluate(\"2018\",\"sf\",\"Tight\",zcands.tag.eta,zcands.tag.pt)\n",
    "sfel2 = evaluator[\"UL-Electron-ID-SF\"].evaluate(\"2018\",\"sf\",\"Loose\",zcands.probe.eta,zcands.probe.pt)\n",
    "\n",
    "sftot = sfel1*sfel2\n",
    "\n",
    "#Apply the weights to the 'Good probes' Z mass peak\n",
    "fig, (axunw, axw) = plt.subplots(1, 2, sharey=True, figsize=(12, 6))\n",
    "mbins = np.linspace(60, 120, 61)\n",
    "\n",
    "axunw.hist(mass[goodprobe], bins=mbins)\n",
    "axunw.set_title('no SF Event weight')\n",
    "axunw.set_ylabel('Events')\n",
    "axunw.set_xlabel('Dielectron mass [GeV]')\n",
    "\n",
    "axw.hist(mass[goodprobe], bins=mbins, weights=sftot[goodprobe])\n",
    "axw.set_title('SF Event weight')\n",
    "axw.set_xlabel('Dielectron mass [GeV]')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Almost imperceptibly different! Look closely at the tails. A ratio plot would be more appropriate.\n",
    "\n",
    "You can of course apply scale factors by the ROOT histograms as well, and a good example of how to do that is included in the [TTGamma Long Exercise](https://twiki.cern.ch/twiki/bin/view/CMS/EgammaUL2016To2018#General_note_about_ID_SFs)! They show show to use the histograms in a Coffea way, but the general principle applies for non-Coffea analyses. This is also a good example of how scale factors can be applied in other scenarios. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "egammaenv",
   "language": "python",
   "name": "egammaenv"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}