Single simulation test on standard reconstruction

This is how we run all single particle productions. Some bugs apparently only affect these simulations. Hopefully this allows us to catch them.

benchmarks/single/analysis/analyze.cxx

+#include <iostream>
+#include <string>
+
+#include <ROOT/RDataFrame.hxx>
+
+#include <eicd/ReconstructedParticleData.h>
+
+int analyze(std::string file)
+{
+  // open dataframe
+  ROOT::RDataFrame df("events", file, {"GeneratedParticles", "ReconstructedParticles"});
+
+  // count total events
+  auto count = df.Count();
+  if (count == 0) {
+    std::cout << "Error: No events found" << std::endl;
+    return -1;
+  }
+
+  auto n_tracks = [](const std::vector<eic::ReconstructedParticleData> &p) { return (int) p.size(); };
+
+  auto d = df
+  .Define("n_tracks_gen", n_tracks, {"GeneratedParticles"})
+  .Define("n_tracks_rec", n_tracks, {"ReconstructedParticles"})
+  ;
+
+  auto stats_n_tracks_gen = d.Stats("n_tracks_gen");
+  auto stats_n_tracks_rec = d.Stats("n_tracks_rec");
+  if (stats_n_tracks_rec->GetMean() < 0.9) {
+    std::cout << "Error: too few tracks per events " << std::endl;
+    stats_n_tracks_gen->Print();
+    stats_n_tracks_rec->Print();
+    return -1;
+  }
+
+  // success
+  return 0;
+}