Project

General

Profile

Bug #25525

Adding offset in MergeSimSources breaks symmetry assumed by truth matching

Added by Edward Tyley 2 months ago. Updated about 2 months ago.

Status:
Assigned
Priority:
Normal
Category:
Simulation
Target version:
-
Start date:
02/17/2021
Due date:
% Done:

0%

Estimated time:
Occurs In:
Experiment:
-
Co-Assignees:
Duration:

Description

When the offset is applied to the TrackID when using the LArSim module MergeSimSources the offset is applied both to the primary track Id and the "rolled up" track Id that is the negative of the original track Id (see https://github.com/LArSoft/larsim/blob/b45eb9ed60d76c792ba445784c457ad048f515fd/larsim/MergeSimSources/MergeSimSources.cxx#L228). For example a muon with Id 1 the delta rays will typically be assigned an Id of -1. When an offset of 1000 is applied these are changed to 1001 and 999 respectively, rather than the 1001 and -1001 one would expect. This breaks a symmetry that is assumed by many truth matching utilities (e.g. https://github.com/LArSoft/larsim/blob/68017a1f0e7f41c444973ae340a003dd41b77aa3/larsim/Utils/TruthMatchUtils.cc#L179) meaning these functions can return Ids that match to non-existent particles.

History

#1 Updated by Lynn Garren 2 months ago

Gianluca, are you able to look at this?

#2 Updated by Kyle Knoepfel 2 months ago

Ed, did you mean to specify LArSoft v06_73_00 in the "Occurs In" field?

#3 Updated by Kyle Knoepfel 2 months ago

  • Status changed from New to Feedback

#4 Updated by Edward Tyley 2 months ago

No sorry, this bug occurred during an SBND production using v09_08_00 but I was not able to select this version from the drop-down menu, and in the process of trying accidentally selected the wrong version.

#5 Updated by Kyle Knoepfel about 2 months ago

  • Assignee set to Gianluca Petrillo
  • Status changed from Feedback to Assigned

Also available in: Atom PDF