/// \file
/// \ingroup tutorial_dataframe
/// \notebook -draw
/// This tutorial illustrates how use the RDataFrame in combination with a
/// RDataSource. In this case we use a TCsvDS. This data source allows to read
/// a CSV file from a RDataFrame.
/// As a result of running this tutorial, we will produce plots of the dimuon
/// spectrum starting from a subset of the CMS collision events of Run2010B.
/// Dataset Reference:
/// McCauley, T. (2014). Dimuon event information derived from the Run2010B
/// public Mu dataset. CERN Open Data Portal.
/// DOI: [10.7483/OPENDATA.CMS.CB8H.MFFA](http://opendata.cern.ch/record/700).
///
/// \macro_code
/// \macro_image
///
/// \date October 2017
/// \author Enric Tejedor

int df014_CSVDataSource()
{
   // Let's first create a RDF that will read from the CSV file.
   // The types of the columns will be automatically inferred.
   auto fileNameUrl = "http://root.cern.ch/files/tutorials/df014_CsvDataSource_MuRun2010B.csv";
   auto fileName = "df014_CsvDataSource_MuRun2010B_cpp.csv";
   if(gSystem->AccessPathName(fileName))
      TFile::Cp(fileNameUrl, fileName);
   auto df = ROOT::RDF::MakeCsvDataFrame(fileName);

   // Now we will apply a first filter based on two columns of the CSV,
   // and we will define a new column that will contain the invariant mass.
   // Note how the new invariant mass column is defined from several other
   // columns that already existed in the CSV file.
   auto filteredEvents =
      df.Filter("Q1 * Q2 == -1")
        .Define("m", "sqrt(pow(E1 + E2, 2) - (pow(px1 + px2, 2) + pow(py1 + py2, 2) + pow(pz1 + pz2, 2)))");

   // Next we create a histogram to hold the invariant mass values and we draw it.
   auto invMass =
      filteredEvents.Histo1D({"invMass", "CMS Opendata: #mu#mu mass;#mu#mu mass [GeV];Events", 512, 2, 110}, "m");

   auto c = new TCanvas();
   c->SetLogx();
   c->SetLogy();
   invMass->DrawClone();

   // We will now produce a plot also for the J/Psi particle. We will plot
   // on the same canvas the full spectrum and the zoom in the J/psi particle.
   // First we will create the full spectrum histogram from the invariant mass
   // column, using a different histogram model than before.
   auto fullSpectrum =
      filteredEvents.Histo1D({"Spectrum", "Subset of CMS Run 2010B;#mu#mu mass [GeV];Events", 1024, 2, 110}, "m");

   // Next we will create the histogram for the J/psi particle, applying first
   // the corresponding cut.
   double jpsiLow = 2.95;
   double jpsiHigh = 3.25;
   auto jpsiCut = [jpsiLow, jpsiHigh](double m) { return m < jpsiHigh && m > jpsiLow; };
   auto jpsi =
      filteredEvents.Filter(jpsiCut, {"m"})
         .Histo1D({"jpsi", "Subset of CMS Run 2010B: J/#psi window;#mu#mu mass [GeV];Events", 128, jpsiLow, jpsiHigh},
                  "m");

   // Finally we draw the two histograms side by side.
   auto dualCanvas = new TCanvas("DualCanvas", "DualCanvas", 800, 512);
   dualCanvas->Divide(2, 1);
   auto leftPad = dualCanvas->cd(1);
   leftPad->SetLogx();
   leftPad->SetLogy();
   fullSpectrum->DrawClone("Hist");
   dualCanvas->cd(2);
   jpsi->DrawClone("HistP");

   return 0;
}