Update menu sequence maker alg (ATR-19824)

In the empty menu sequence needed for merging combined chains serially, we need an Input Maker to merge legs from different filters. The empty sequence is only made of one algorithm, the IM, which is then connected to the ComboHypo.

Currently the IM is the AthenaConfiguration.ComponentFactory.CompFactory.HLTTest__TestInputMaker but should be switched for DecisionHandling.DecisionHandlingConf.InputMakerForRoI

@markowen @tamartin @fpastore

Errors in the trivial change when running test_trigUpgr_emu_step_processing_build.sh with the empty sequences.

IMstep2EmptySeqence                        2   0   DEBUG Executing IMstep2EmptySeqence...
IMstep2EmptySeqence                        2   0   DEBUG Creating one merged output per Decision object
IMstep2EmptySeqence                        2   0   DEBUG Running on input HLTNav_FStep2_mu_em_serial__ComboHypo_Ste
p1_mu_em_serial_v1Step1MuHypo with 3 elements
IMstep2EmptySeqence                        2   0   DEBUG  -- Input Decision 0: has 1 previous links
IMstep2EmptySeqence                        2   0   DEBUG   -- Did not match to existing, HLTNav_FStep2_mu_em_serial
__ComboHypo_Step1_mu_em_serial_v1Step1MuHypo index 0 creates output index 0
IMstep2EmptySeqence                        2   0   DEBUG       -- This output decision now has 1 seeds and 1 decisi
onIds
IMstep2EmptySeqence                        2   0   DEBUG  -- Input Decision 1: has 1 previous links
IMstep2EmptySeqence                        2   0   DEBUG   -- Did not match to existing, HLTNav_FStep2_mu_em_serial
__ComboHypo_Step1_mu_em_serial_v1Step1MuHypo index 1 creates output index 1
IMstep2EmptySeqence                        2   0   DEBUG       -- This output decision now has 1 seeds and 1 decisi
onIds
IMstep2EmptySeqence                        2   0   DEBUG  -- Input Decision 2: has 1 previous links
IMstep2EmptySeqence                        2   0   DEBUG   -- Did not match to existing, HLTNav_FStep2_mu_em_serial
__ComboHypo_Step1_mu_em_serial_v1Step1MuHypo index 2 creates output index 2
IMstep2EmptySeqence                        2   0   DEBUG       -- This output decision now has 1 seeds and 1 decisi
onIds
IMstep2EmptySeqence                        2   0   DEBUG Got output HLTNav_IMstep2EmptySeqence__out with 3 elements
IMstep2EmptySeqence                        2   0   DEBUG Found RoI: RoIid: 0 RoIword: 0  z: 0 (-225 - 225) eta: 2 (
1.9 - 2.1) phi: 0.5 (0.4 - 0.6) FS=0
IMstep2EmptySeqence                        2   0   DEBUG Added RoI: RoIid: 0 RoIword: 0  z: 0 (-225 - 225) eta: 2 (
1.9 - 2.1) phi: 0.5 (0.4 - 0.6) FS=0
IMstep2EmptySeqence                        2   0   DEBUG Found RoI: RoIid: 0 RoIword: 0  z: 0 (-225 - 225) eta: 3 (
2.9 - 3.1) phi: 0.5 (0.4 - 0.6) FS=0
IMstep2EmptySeqence                        2   0   DEBUG Added RoI: RoIid: 0 RoIword: 0  z: 0 (-225 - 225) eta: 3 (
2.9 - 3.1) phi: 0.5 (0.4 - 0.6) FS=0
IMstep2EmptySeqence                        2   0   DEBUG Found RoI: RoIid: 0 RoIword: 0  z: 0 (-225 - 225) eta: 2.2
 (2.1 - 2.3) phi: 0.6 (0.5 - 0.7) FS=0
IMstep2EmptySeqence                        2   0   DEBUG Added RoI: RoIid: 0 RoIword: 0  z: 0 (-225 - 225) eta: 2.2
 (2.1 - 2.3) phi: 0.6 (0.5 - 0.7) FS=0
IMstep2EmptySeqence                        2   0   DEBUG Produced 3 output RoIs
0_StoreGateSvc_Impl                        2   0 WARNING  setupProxy:: error setting up proxy for key Unspecified a
nd clid 1097199488
 Pre-existing valid DataProxy @0x18d32640 found in Store for key Unspecified with clid 1097199488
0_StoreGateSvc_Impl                        2   0 WARNING record_impl: Problem setting up the proxy for object @0x1a
ab2fc0
 recorded with key Unspecified of type TrigRoiDescriptorCollection (CLID 1097199488) in DataObject @0x1a32a8c0
VarHandle(StoreGateSvc+Unspecified[10...   2   0   ERROR StoreGate/src/VarHandleBase.cxx:737 (StatusCode SG::VarHan
dleBase::record_impl(std::unique_ptr<DataObject>, void*, bool, bool)): code FAILURE: recordObject failed
IMstep2EmptySeqence                        2   0   ERROR DecisionHandling/src/InputMakerForRoI.cxx:76 (StatusCode I
nputMakerForRoI::execute(const EventContext&) const): code FAILURE: roi_outputHandle.record(std::move(oneRoIColl))
ExceptionSvc                               2   0   DEBUG Service base class initialized successfully
ExceptionSvc                               2   0   DEBUG Handling Error from IMstep2EmptySeqence
IMstep2EmptySeqence                        2   0   ERROR Maximum number of errors ( 'ErrorMax':1) reached.
AlgoExecutionTask                          2   0 WARNING Execution of algorithm IMstep2EmptySeqence failed

Hi @vcroft

Right, we don't care about writing an ROI collection here - there is no reco sequence which can use it as a concrete starting point. We can relax the requirement on a ROI WriteHandle key.

(To me, making the writing of the ROIs optional is better than adding yet another InputMaker class type)

Opening quick MR - you could also merge this into your branch.

mentioned in merge request !31672 (closed)

added 371 commits

• 04b0f7ba...99544431 - 369 commits from branch atlas:master
• 84d9b2d6 - Make ROI collection output optional
• ca013232 - Merge tims IM branch 'tim/inputMakerAllowEmpty' into empty-sequence-maker

Compare with previous version

Hi @tamartin thanks for the quick response. Is there anything I should do other than merge? - I'm still getting errors.

Hm, assuming you added Trigger/TrigSteer/DecisionHandling to your build, I thought that would do it. Same error?

resolved all threads

Hi @fpastore @tamartin new errors here with regards to the handling of MuonHypos with an experimental serial chain in the physics menu.

TrigL2MufastHypoAlg                        5   0   ERROR TrigMuonHypoMT/src/TrigMufastHypoAlg.cxx:81 (StatusCode Tr
igMufastHypoAlg::execute(const EventContext&) const): code FAILURE: forIDHandle.isValid()
...
TrigL2MufastHypoAlg (285), w/ decision: UNDEFINED(-1), in state: ERROR
ComboHypo_merged_Step2_Step1_HLT_e3_etcut1step_mu6fast_L1EM8I_MU10EmptySeq1_1_Step1_1mufast (286), w/ decision: UNDEFINED(-1), in state: INITIAL

I'm not sure if this is relevant but, since Tims MR of this morning included the RDO to PRD with PRD caching (CSC, MDT) I'm now building this against all the packages listed in that MR

+ MuonSpectrometer/MuonCnv/MuonCSC_CnvTools
+ MuonSpectrometer/MuonCnv/MuonMDT_CnvTools
+ MuonSpectrometer/MuonCnv/MuonMM_CnvTools
+ MuonSpectrometer/MuonCnv/MuonRPC_CnvTools
+ MuonSpectrometer/MuonCnv/MuonRdoToPrepData
+ MuonSpectrometer/MuonCnv/MuonSTGC_CnvTools
+ MuonSpectrometer/MuonCnv/MuonTGC_CnvTools
+ MuonSpectrometer/MuonConfig
+ MuonSpectrometer/MuonReconstruction/MuonRecEvent/MuonPrepRawData
+ MuonSpectrometer/MuonReconstruction/MuonRecEvent/MuonTrigCoinData
+ Trigger/TrigSteer/DecisionHandling
+ Trigger/TriggerCommon/TriggerJobOpts

any ideas?

That change (to remove the redundant "forID" ROI link) was in !31510 (merged) which went in yesterday afternoon.

If you try compiling against r31 (which only became available this afternoon) then I would hope you don't see it. Compiling against r28 you should also add

+ Trigger/TrigHypothesis/TrigMuonHypoMT
+ Trigger/TrigSteer/ViewAlgs
+ Trigger/TriggerCommon/TriggerMenuMT

resolved all threads

Ok I think everything is fine now I just need to do some minor cleanups to make sure this works against the new master as is then I'll un WIP it and set it going into the wild.

unmarked as a Work In Progress

changed title from WIP: update menu sequence maker alg (ATR-19824) to Update menu sequence maker alg (ATR-19824)

changed the description

This merge request affects 3 packages:

• Trigger/TrigSteer/DecisionHandling
• Trigger/TrigValidation/TrigUpgradeTest
• Trigger/TriggerCommon/TriggerMenuMT

Adding @tamartin ,@hartj ,@jpanduro as watchers

added Trigger TriggerMenu master review-pending-level-1 labels

CI Result SUCCESS (hash ca013232)

	Athena	AthSimulation	AnalysisBase
externals
cmake
make
required tests
optional tests

Full details available on this CI monitor view
Athena: number of compilation errors 0, warnings 0
AthSimulation: number of compilation errors 0, warnings 0
AnalysisBase: number of compilation errors 0, warnings 0
For experts only: Jenkins output [CI-MERGE-REQUEST-CC7 11802]

Push to L2 since there is a lot of changes.

Xiaozhong (L1)

added review-pending-level-2 label and removed review-pending-level-1 label

@vcroft - the list of changes looks a bit weird to me - it seems to include things we already merged in !29830 (merged).

Might be connected with using the same source branch?

resolved all threads

Changes looks fine from shifter point of view, also expert questions were answered, so approving.

Pavol [as L2 MR shifter]

added review-approved label and removed review-pending-level-2 label

merged

mentioned in commit 92ca3df7

added sweep:ignore label