mCBM 2020: Event based reco with STS, RICH and PSD
- Fix name of event array in mCBM related classes
- Add RICH and PSD reconstruction to the two event based macros
- Clean and re-organize the macros to remove obsolete code and make it easier to compare them
Related to redmine issue refs #1827
Merge request reports
Activity
- Resolved by Pierre-Alain Loizeau
@karpushkin_AT_inr.ru, @a.weber: I modified some of your mCBM reco classes to make them compatible with the new name of the TClonesArray (and with the generic name from simulation ;-) ), please have a look at the MR changes for your folders and let me know if this is OK by approving the Merge Request or let me a comment if I should revert it.
@v.singhal and @n.herrmann : The reconstruction code for MUCH and TOF was already compatible with both option. Please let me know if I should cleanup the old mCBM one in the following files (by default I will leave it).
Double support in MUCH code: reco/detectors/much/CbmMuchFindHitsGem.cxx 95-97
Double support in TOF code: reco/detectors/tof/CbmTofEventClusterizer.cxx 525-526 reco/detectors/tof/CbmTofFindTracks.cxx 234-235 analysis/detectors/tof/CbmTofAnaTestbeam.cxx 949-950 analysis/PWGHAD/hadron/CbmHadronAnalysis.cxx 2929-2930
Unsure: MQ/hitbuilder/CbmDeviceHitBuilderTof.cxx 386
For all of you and @a.toia, I also
- modified the reco macros to make them consistent between each other
- introduced the RICH and PSD blocks in the event based ones (mcbm_event_reco and mcbm_build_and_reco_kronos)
- Enabled the event mode for STS in the two same macros
=> Please let me know if this broke something for you, in which case I will try to correct before we merge
Unfortunately it will still take a few weeks to get the automatic tests on all those macros so for now we have to rely on human feedback.
@p.-a.loizeau: All changes look fine for me. I don not see any problem :)
added 1 commit
- ec4dba2d - mCBM 2018-2020: improve handling event array retrieval in EventCheck
The change from
CbmEvent
toEvent
for the TClonesArray name actually brought down a problem: we now have a name collision between the name of the Array and the TFolder in which FairRoot stores the FairEventHeader.This confuses the
GetObject
method of the FairRootManager, which simply returns the first object it finds, in this case the TFolder, which in turn leads to the failed test.You can see this in the following screenshot (FairEventHeader in the unpack file declared as input and CbmEvent in the event file declared as Friend):
Most probably the straightforward solution would be to either
- Make an exception to the rule for
CbmEvent
as Event is too generic, and use CbmEvent as Array name everywhere (needs to change in all simulation related code) - Rename
CbmEvent
to something else so the rules still apply while giving a unique name (also need a change everywhere)
Should we discuss this in tomorrow meeting if there is time?
- Make an exception to the rule for
Under these circumstances, the best seems to me to let the branch name be CbmEvent in this case and do the proper changes in the reconstruction tasks (not in the simulation, I think; this branch is not created in the simulation). I do not think a discussion in the meeting is needed or would be helpful.
Ok.
I will try to prepare replacement MR with the necessary changes to the reco and tracking code (that was indeed a late afternoon mixup by me to say simulation).
Should I introduce a similar switch as in TOF and MUCH to provide support of both names (
Event
andCbmEvent
) avoid backward compatibility issues with already generated reco files?added 9 commits
-
ec4dba2d...55b769bf - 5 commits from branch
computing:master
- f5922b23 - mCBM 2020: add PSD and RICH to event based reco + clean/synch the macros
- 20c4b9c9 - Use dynamic cast when getting the pointer on CbmEvent array
- 8caf420a - Unify name of CbmEvent array to CbmEvent to avoid name collision
- e0816a58 - Apply format
Toggle commit list-
ec4dba2d...55b769bf - 5 commits from branch
added 1 commit
- 0e00ac70 - Change name of CbmEvent array in reco classes in mvd folder
Changing assignment to @v.friese as we went from modifications in the mCBM code to lots of small modifications in the reco code.
assigned to @v.friese and unassigned @p.-a.loizeau
added 18 commits
-
0e00ac70...7f6a2088 - 13 commits from branch
computing:master
- 35ce4afb - mCBM 2020: add PSD and RICH to event based reco + clean/synch the macros
- 827b84d8 - Use dynamic cast when getting the pointer on CbmEvent array
- e8fdf9e2 - Unify name of CbmEvent array to CbmEvent to avoid name collision
- db9175d2 - Apply format
- f26bc3eb - Change name of CbmEvent array in reco classes in mvd folder
Toggle commit list-
0e00ac70...7f6a2088 - 13 commits from branch
added 6 commits
-
ce8f144c - 1 commit from branch
computing:master
- a2102a2b - mCBM 2020: add PSD and RICH to event based reco + clean/synch the macros
- 6c772d25 - Use dynamic cast when getting the pointer on CbmEvent array
- 6864e354 - Unify name of CbmEvent array to CbmEvent to avoid name collision
- d7b8325d - Apply format
- b24b5a16 - Change name of CbmEvent array in reco classes in mvd folder
Toggle commit list-
ce8f144c - 1 commit from branch