hep-fcc / fccdetectors Goto Github PK

Some detector source / compact files are currently present in triplicate in the key4hep-stack - in DetFCCeeIDEA, DetFCCeeCLD and LCGEO. Since historically the CLD detector model was in LCGEO, I'd propose to keep the files there and only keep a master .xml in this repo that uses $LCGEO for includes.

Attempting to run PandoraPFA inside DDMarlinPandora inside k4MarlinWrapper with the following steering instructions:

from Configurables import MarlinProcessorWrapper

pandora = MarlinProcessorWrapper('DDMarlinPandora')
pandora.OutputLevel = DEBUG
pandora.ProcessorType = 'DDPandoraPFANewProcessor'
pandora.Parameters = {
    'Verbosity': ['WARNING'],
    'PandoraSettingsXmlFile': ['/some/path'],
    'CreateGaps': [False],
    'ECalCaloHitCollections': ['ECalBarrelCells']
}
ApplicationMgr().TopAlg += [pandora]

results in the following error:

[ ERROR "DDMarlinPandora"] Failed to initialize marlin pandora: std exception  getExtension: selection is not unique (or empty)  includeFlag: DetType( 0x812) : CALORIMETER, BARREL, ELECTROMAGNETIC,  excludeFlag: DetType( 0x20020) : FORWARD, AUXILIARY,  --- found detectors : 

Full steering macro is in this gist and output log are in the logfile.txt.

see AIDASoft/DD4hep#817

I think I found a bug in Detector/DetSegmentation/include/DetSegmentation/GridTheta.h in the method positionFromRThetaPhi(double ar, double atheta, double aphi) const

The code calculates z as if the input were spherical coordinates, but actually other parts of the code call this method passing to it cylindrical coordinates.
This leads to mismatch in theta between the positioned hits and the corresponding position cells.

So, the buggy line is
return Vector3D(ar std::cos(aphi), ar std::sin(aphi), ar * std::cos(atheta));

the correct code that should replace it is:
return Vector3D(ar std::cos(aphi), ar std::sin(aphi), ar / std::tan(atheta));

(It seems that one should use cylindrical coordinates as one could infer also from GridEta (and since this seems to be more natural given the typical cylindrical symmetry of our detectors.)

...should be made configurable.

From the name of the model FCCee_o2_v02 I would assume that this is based on FCCee_o2_v01, which is the CLD detector model with the smaller 10 mm beampipe, and the corresponding smaller vertex detector
cf. https://github.com/iLCSoft/lcgeo/blob/4a5b6bb293ce5526ed23d14a3b5157a31326caf8/FCCee/compact/FCCee_o2_v01/FCCee_o2_v01.xml#L284

But looking at the vertex detector used in FCC_o2_v02 I see:

Instead of
https://github.com/iLCSoft/lcgeo/blob/4a5b6bb293ce5526ed23d14a3b5157a31326caf8/FCCee/compact/FCCee_o2_v01/Vertex_o4_v05_smallBP.xml#L41-L43

	<constant name="VertexBarrel_r1" value="1.25*cm"/> 
        <constant name="VertexBarrel_r2" value="3.5*cm"/>
        <constant name="VertexBarrel_r3" value="5.7*cm"/>

Hi, I tried to run ddsim on the CLD compact files but it does not work.
I am sourcing the nightlies and trying to use $FCCDETECTORS/Detector/DetFCCeeCLD/compact/FCCee_o2_v02/FCCee_o2_v02.xml

Is this expected?
ddsim works with ILC compacts from $lcgeo_DIR

I wanted to make the subdetectors which have their own subdirectories more standalone, and started using the elements.xml from DD4hep for those. But it might be a good idea in general to pick this file up from DD4hep.