Minutes for LArSoft coordination meeting on May 21, 2019

Present | Lynn Garren, Krzysztof Genser,Robert Hatcher, Alex Himmel, Kyle Knoepfel, Saba Sehrish, Erica Snider, Hans Wenzel|

Remote | Katherine Lato |

Release and Project status report [Erica Snider]

• 08_20_00 updates to pandora, bug fixes, improvement to OpDet
• this week, ifdhc update
  • recob::Vertex fix is already in develop
• migration to Python 3
  • means dropping slf6
  • We need to know if it is supported by experiments, and in what timeframe?
  • LArIAT has alot of Python scripts, who is going to update those. Its experiments’ responsibility.
  • Infrastructure things are already in Python 3, FIFE is pushing for end of summer migration
  • project.py in LArbatch
• nutools refactoring
  • version tightly coupled with specific versions of Geant4, Genie,
  • re-factor into separate components
  • plan is to have a proposal in the next LCM
  • dedicated discussion on Thursday

Conclusion

• Need to know from experiments about migration to Python 3
  • end of life for Python 2 in January so we should try no later than end of summer deadline

Survey of module and service usage across LArSoft [Kyle Knoepfel]

• Survey was sent April 29
• The main goal is to remove the code no one is using
• 18 people from multiple experiments responded, comment may not have enough coverage
• The slide shown had a service with votes: “use” or “remmove”. green color is use it, blue is remove it
• Questions asked in the survey were: What services do you use, and which ones should be removed?
  • same for filters and analysers
• AtomicNumberService, ShowerCalibrationGaloreFromPIDService, ShowerCalibrationGaloreScaleService are in larexamples
  • ask why it was marked for removal
• larreco/Deprecated directory should be removed
• Anything that has the name cheater in it is used
• Overall there is an agreement on the 2-step proposal presented, there will be follow-up during the upcoming coordination meetings

• email will be sent out with the list of services and modules to remove

• A side announcement: ICARUS will be using some experiments’ (may be DUNE’s) disambiguation code

New features and interface changes [Gianluca Petrillo]

• In Geant4 wire plane is a box and has thickness, it has wires that are cylinders
  • 2 or 3 wire planes are adjacent to the others
• if you ask for the position of wires, LArG4 voxel readout was returning the center
  • geo::TPCGeo::PlaneLocation() reports the center of the box
  • where wires are not in the center, in case of Argoneut the wires were on the side
• consider geo::TPCGeo::PlaneLocation() deprecated; there are better options:
  • geo::PlaneGeo::GetCenter(): less fragile
  • geo::TPCGeo::DistanceToReferencePlane(), geo::PlaneGeo::DriftPoint(),
  • geo::PlaneGeo::DistanceFromPlane(): drift-direction aware
• Request to officially deprecate it
  • does it make sense right now? provide feature branches with deprecation
• discussion to clear out the different concepts in readout, simulation etc

• comment on if utilities are not widely used across all the experiments, then they should not be added to LArSoft

• Comment by Krzysztof Genser on naming conventions that LArSoft should have one.

• Geant4 has the naming convention to use G4, so other projects should not use classes named that way. G4Helper e.g. in nutools