karuboniru / rdf_framework

Automatic QUality Assurance (AQUA)

TL;DR:

This tools can load a set of ROOT file, do user defined analysis and produce plots and TTrees for further plotting and those plots can be merged via this tool to make parallel comparison.

Configure file format

This is the compoments of a configuration file, for more details please check the example file in example/ folder.

{
  "files": [
    "/dybfs/users/chengjie/data/GENIE/GENIE3.0.6-wDeex/useroutput_wdeex.root",
    "/path/contains/many/root/files/*.root",
    "/path/contains/many/root/files/*/output.root"
  ],
  "treename": "atmgen", // tree name in the root file to send to analysis code
  "plot_file": "plot.root", // output file name to put the TH1 in 
  "preprocess": "CC_selection", // Optional, function name in the plugin, preprocess the input tree before sending to analysis code, this can include global cuts or global variable to be used
  "normalize": {  // Optional, if this block is given, the output TH1 will be normalized to the given value, using `Scale(norm, "WIDTH");`
    "func": "normalize_factor_CC", // func name given in the plugin file
    "parameters": [ // a list of parameters to be passed to the function
      "/media/storage/neutrino/GENIE/atmo/G18_10b_02_11b.root",
      "nu_mu_C12/tot_cc",
      "12"
    ]
  },  
  "plugins": [
    {
      "name": "Generator",  // name of the plugin
      "plugin": "libmuon_p.so", // plugin file path, use full path if nessary,
      "analysis": [
        {
          "treename": "treeout", // tree name to put the output tree in
          "filename": "muon_kin.root", // output file name to put the output tree in
          "func": "analysis", // function name in the plugin
          "output": [
            "muon_p",
            "muon_e",
            "leading_muon_e",
            "leading_muon_p" // variables to be captured in the output tree, can be variables found in input tree or variables calculated in the analysis function
          ],
          "plots": [
            {
              "var": "leading_muon_p",  // var to be plotted
              "name": "leading_muon_p", // name of the plot saved
              "xmin": 0,
              "xmax": 0,
              "bins": 128,
              "cut": "leading_muon_p>0.2" // cut can be formular to be evaled or func name of cut given in plugin file
            },
            {
              // can be more variables to plot
            }
          ],
          "stack_plots": [
            {
              "var": "dalphat",
              "name": "dalphat",
              "xmin": 0,
              "xmax": 0,
              "bins": 128,
              "cuts": [  //currently only support stacking different cuts on a same variable, can be more than 2 cuts
                "W < 1.4",
                "W > 1.4"
              ]
            }
          ]
        }
      ]
    }
  ]
}

Analysis plugin

The analysis plugin is a shared library that can be build standalone or with the help of cmake. The plugin should contain several function with the following signature:

class Name : ProcessNodeI{
public:
  ROOT::RDF::RNode operator()(ROOT::RDF::RNode df) override {
    return df/*whatever*/;
  }
  void configure (const nlohmann::json &conf) override {
    /* do read information from config */
  }
};
REGISTER_PROCESS_NODE(Name)

For a normalize function, the signature should be:

class NAME : public NormalizeI {
public:
  double operator()(ROOT::RDF::RNode df) override {
    // return normalization factor
  }

  void configure(const nlohmann::json &conf) override {
  }
};
REGISTER_NORMALIZE(NAME)

All cuts and analysis is done in the way of mutating the RNode and return it. You can find documentation of RNode here. Or read given code in example/muon_p.cxx and its comment.

To give configuration, when specifying "func": "analysis" do

"func": {
  "name": "normalize_factor_cc",
  "config": {
    "filename": "/media/storage/neutrino/GENIE/atmo/G18_10b_02_11b.root",
    "obj_path":"nu_mu_C12/tot_cc",
    "Z": 12
  }
}

instead, whatever in "config" will be passed to configure before the actual mutation is called.