[Rivet] Rivet analysis ATLAS_2010_S8914702

[Holger Schulz]

Hi Michael,

during testing we found that your analysis plugin crashes for certain events
when trying to calculate the median pTDensity.

Could I please ask you to have a look at these code snippets.


       // Get the jet pT densities
       vector< vector<double> > ptDensities(_eta_bins_areaoffset.size()-1);
       FastJets fastjets = apply<FastJets>(event, "KtJetsD05");
       const shared_ptr<fastjet::ClusterSequenceArea> clust_seq_area = 
fastjets.clusterSeqArea();
       for (const Jet& jet : fastjets.jets()) {
         const double area = clust_seq_area->area(jet); //< Implicit 
call to pseudojet()
*/// @todo Should be 1e-4?*
         if (area > *10e-4* && jet.abseta() < _eta_bins_areaoffset.back()) {
           ptDensities.at(getEtaBin(jet.abseta(), true)) += jet.pT()/area;
         }
       }

       // Now compute the median energy densities
       vector<double> ptDensity;
       for (size_t b = 0; b < _eta_bins_areaoffset.size()-1; ++b) {
*if (ptDensities[b].size() >0 )* ptDensity += median(ptDensities[b]);
       }

The red bit was added by us to prevent calculating the median of an 
empty vector --- do you remember
what the logic in your analysis code was for cases like that?

If you find the time, could you please comment on the blue bit as well?

You can find the version of the code in the current release here:

    https://rivet.hepforge.org/hg/rivet/file/0175188f89df/src/Analyses/ATLAS_2010_S8914702.cc


Cheers,
Holger

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.hepforge.org/lists-archive/rivet/attachments/20160720/77e30154/attachment.html>