Geant4 plugin

The Geant4 plugin allows to build Acts::TrackingGeometry and Acts::Experimental::Detector directly from Geant4 geometry input. Both rely on the conversion of G4VPhysicalVolume into corresponding Acts objects.

Object conversion

Surface conversion

Converting physical volumes into Acts::Surface objects that represent sensitive detector elements, is done via the Acts::Geant4DetectorSurfaceFactory. This helper class allows to select volumes from the Geant4 geometry and convert them either into pairs of Acts::Geant4DetectorElement and Acts::Surface objects in case of sensitive elements, or simply surfaces objects in the case of passive surfaces.

The selection is hereby done by providing one or more Acts::IGeant4PhysicalVolumeSelector objects to the surface factory.

Possible implementations of this type of conversions can be seen in the corresponding unit test ActsUnitTestGeant4DetectorSurfaceFactory

  // Get the box
  auto nameSelector =
          std::vector<std::string>{"yl"}, false);

  Acts::Geant4DetectorSurfaceFactory::Cache cache;
  Acts::Geant4DetectorSurfaceFactory::Options options;
  options.sensitiveSurfaceSelector = nameSelector;

  Acts::Geant4DetectorSurfaceFactory factory;
  factory.construct(cache, nominal, *cylinderPV, options);

  BOOST_CHECK_EQUAL(cache.sensitiveSurfaces.size(), 1u);
  BOOST_CHECK_EQUAL(cache.passiveSurfaces.size(), 0u);

  auto [ element, surface ] = cache.sensitiveSurfaces.front();
  BOOST_CHECK_EQUAL(surface->type(), Acts::Surface::SurfaceType::Cylinder);

Inspecting surface conversion within python

The ActsExamples python bindings allow to conveniently test the conversion of Geant4 volumes into sensitive and passive surfaces, assuming you have a GDML file called detector.gdml where Geant4PhysVolume objects can be identified by a certain string, e.g. names containing the flag Sensitive, or Passive. Also, multiple match strings are allowed. The converted surfaces can then be displayed with .obj (part of the Core functionality) or as .svg files (if ACTS_BUILD_PLUGIN_ACTSVG is switched on)

  # import the necessary modules
  import acts, acts.examples
  from acts.examples import geant4 as acts_g4
  # The match criteria
  sensitive_matches = [ 'Sensitive' ]
  passive_matches = [ 'Passive' ]
  [ elements, ssurfaces, psurfaces ] = acts_g4.convertSurfaces('detector.gdml', sensitive_matches, passive_matches)

  # Some screen output
  print('* Conversion yielded', len(ssurfaces))

  # Write them to an obj file 
  drawContext = acts.GeometryContext()
  sensitiveRgb = [ 0, 150, 150 ]
  passiveRgb = [ 150, 150, 0]
  # Draw the sensitive surfaces
  acts.examples.writeSurfacesObj(ssurfaces, drawContext, sensitiveRgb, 'detector-sensitives.obj')
  # Draw the passive surfaces
  acts.examples.writeSurfacesObj(psurfaces, drawContext, passiveRgb, 'detector-passives.obj')

Building a Detector from Geant4 input

In order to build an Acts::Detector object from Geant4 input, the following steps needs to be done

  • a conversion of Geant4PhysVolume objects into Acts::Surface and Acts::DetectorVolume objects (see before)

  • a build sequence needs to be defined and the converted objects identified

There are several helper methods and tools that can be used, many of them accessible through python bindings. One core component is the selection and assignment of surfaces to dedicated volume. This can be done using e.g. a KDT structure, this can be tested with:

 # Create a KDTree from all surfaces binned in z and r
 surfacesKdt = acts.KdtSurfaces2D(buildContext, surfaces, [acts.Binning.z, acts.Binning.r])

 # Define a query range and select
 qrange = acts.Range2D( [-580,580], [0,200])
 selected = surfacesKdt.surfaces(qrange)

 # Draw, inspect the surfaces

Selected surfaces can be put as a layer structure into a volume