Trying to fix the PRD collection names in the Run 2 muon trigger configuration. We can't currently use the same collection names in the trigger and offline when running the offline and trigger in the same jobs (we run into the problems reported in ATR-20971).
However, there are many tools that depend on these collection names, and it's not that straight forward to change them all (and for code that will be obsolete soon, isn't really worth the effort). Since this is only a problem for the Run 2 trigger code, and this is only used for validation comparisons between Run 2 and Run 3 triggers where we don't run the trigger and offline reco in the same job, will do the following:
- Use the default offline names in jobs where we are not going to be running the offline and trigger at the same time, so we get correct Run2 trgger results
- Will use different names when we are running the trigger and offline at the same time and accept that the HLT muon trigger output is not 100% correct (it's a small effect anyway)

This reverts merge request !30589

Updating the decoding tools used in Run 2 trigger to output to default collections (since these are used in quite a lot of places later on, and the collection names don't need to be different from the default).
Also no longer ignoring some misalgined CSC chambers as we did for Run 2 to make it easier to compare to the Run 3 triggers where we won't ignore these chambers by default.

Nicolas Koehler authored 5 years ago and Walter Lampl committed 5 years ago

Remove circular dependency between CombinedMuonTrackBuilder, TrackQuery and TrackCleaner (ATLASRECTS-5272)

Since the Run 2 version of the muon trigger still requires the non-MT version of the BS decoding, updating the configuration to explicitly use those tools.
Fixes ATR-20906.

Removing the functionality to calculate the muon trigger isolation using FTK tracks and L2StandaloneMuons from the muon trigger isolation tool and associated algorithms (ATR-20766).
Also removed some obsolete functions leftover from Run 1 (using TrigMuonEFInfo object), and removed some unneeded include statements.

Removing the FTK muon isolation also from the menu code (and removed the other muon related FTK chains from the physics_pp_v7_primaries menu), so that no one tries to use this when it isn't supported.

- print function.

Changing the names of the output containers (and relevent input containers) for the RDO to PRD decoding for the Run 2 muon triggers. This is to avoid clashes in trying to write to the same containers when running the offline reconstruction and the legacy run 2 triggers at the same time.
Since the L2MuonSA config is used for both run 2 and run 3 triggers, this change only applies when we run in the run 2 mode (for doMT=False). The TrigMuonEF configuration is only used for the run 2 trigger case, so the changes are applied directly.
Fixes ATR-20615

Savanna Shaw authored 5 years ago and Vakhtang Tsulaia committed 5 years ago

Since the caching was disabled in the MT version of the RDO to PRD decoding tools, we should use the non-MT version in the Run 2 triggers. Otherwise we will run into problems similar to ATR-20615 when we try to run in events with more than one RoI.

Goetz Gaycken authored 5 years ago and Walter Lampl committed 5 years ago

- Use a PRD-to-track map rather than the PRD_AssociationTool which conains a PRD-to-track map
  internally. This is needed to resolve ATLASRECTS-5005.
- Made all methods const.
- The score processor does not create track copies anymore, thus PRD-to-track maps stay valid for
  scored tracks.
- The ambiguity processors now use a helper class which manages the new and original tracks i.e.
  tracks with mixed ownership, and creates track copies only for the final track collection if
  necessary.
- Moved debug ntuple code of simple ambiguity processor to a new derived class, to improve the
  readability.

Dropped the redundant callback declaration and the redundant BuildVolumesFromTagInfo property
(which has never been set to True). Cleaned up the ITrackingVolumesSvc interface and the client
code.

William Axel Leight authored 5 years ago and Frank Winklmeier committed 5 years ago

The service now uses the MdtCondDbData object created by the MdtCondDbAlg, rather than the MdtCondSummarySvc.
The MdtIntersectGeometry objects are no longer cached in the service but created on the fly as needed, making the service much simpler.
Mutable data members are removed or made local variables.
The MdtCondDbData object and the MuonDetectorManager must now be passed in from outside: the clients of the service are adjusted accordingly.
The MdtCondDbAlg must now run during muon reconstruction: the configuration is updated accordingly.

William Axel Leight authored 5 years ago and Edward Moyse committed 5 years ago

Mutable variables are moved onto the stack and the code is migrated to use smart pointers.
The tool now returns a std::unique_ptr: calling tools are modified to expect this.
As this makes it impossible to compare pointers to see if there have been any changes from the cleaning, track parameters are now compared instead.
Finally, the TMEF_CombinedMuonTrackBuilder will no longer add a perigee to the track at the MS entrance, matching the configuration in the offline tool.
This also avoids possible problems resulting from creating a copy of a Trk::Track object with two perigees.

Edward Moyse authored 5 years ago and Walter Lampl committed 5 years ago

Migrated a bunch of Muon tools to be private. For some tools, this is not possible since they contain state. Some configuration changes too to clarify some probable bugs.

The actual hole search is performed by the ID helper tool. The only
purpose of the tool was to toogle the ID hole search on or off.
The latter is now handled by a boolean property as is the case
for muons.

We never actually use the MuidCaloEnergyMeas tool in the trigger, since we don't have access to calo information at the muon trigger level. Any algorithms/tools that need it we anyways configure such that this tool is disabled.
Removing this from the trigger configuration so some cleanup of the offline code can proceed in atlas/athena!23065.

Savanna Shaw authored 5 years ago and Adam Edward Barton committed 5 years ago

Includes proper configuration of EF muon isolation algorithm and hypo,
setup of proper sequence to run ID tracking around muons and calculate the
isoltion, and addition of the relevant pieces to the menu.

In preparation for the migration of the Athena code to the multi-threaded AthenaMT, remove the (scheduler-unsafe) dependency of the TGC RDO->PRD tool on the existence of the RDOs which are in some cases produced from the BS using a decoding function called by the RDO-PRD process itself.

The RawDataProviderTool which can perform the BS decoding is removed from the Muon Spectrometer TGC RDO->PRD convenience tool and can now be used directly by trigger and offline in a similar manner. Changes have been made to the tigger validation unit tests to reflect these changes.

Update packages:TrigmuComb TrigmuIso TrigMuonEF TrigmuRoI TrigMuSuperEF TrigTileMuId TrigMuonCoinHierarchy TrigMuonMatching muon  from upstream/21.0-TrigMC to upstream/master via pseudo-merge

Neil Warrack authored 6 years ago and Vakhtang Tsulaia committed 6 years ago

Former-commit-id: 35816d61

Changes made:
- TrigMuonEFStandaloneTrackTool: Add new job property DecodeBS to control whether or not to do the BS->RDO decoding.
- TrigMuonEFConfig: Configure the DecodeBS flag based on DetFlags and make sure to use the correct RpcRawDataProviderTool


Former-commit-id: 46edac6a4929cd636e56d189f4bcd378c9995425

Remove the conversion of the BS to RDO from the Muon Spectrometer code (RpcRdoToPrepDataTool.cxx) and replace it in the HLT code (TrigMuonEFStandaloneTrackTool.cxx AND RpcDataPreparator.cxx).
Make corresponding edits on header files.
Remove reference to the BS conversion tool in the python scripts MuonPrdProviderToolsConfig.py and TrigMuonEFConfig.py.


Former-commit-id: 0a6d57d5a54b187e5566110ecd9779a1cef4814b

Andrea Coccaro authored 7 years ago and Atlas Nightlybuild committed 7 years ago

Clean up of muon trigger online histograms (ATR-17777)

See merge request atlas/athena!9567

(cherry picked from commit d0e4d7f1 [formerly cf47fdb4])

aecd8268 Clean up of muon trigger online histograms (ATR-17777)

Former-commit-id: 0a72f921

Former-commit-id: 9b39c6d9

The MooSegmentCombinationFinder property "WriteIntermediateSegments", which was always false, has been removed, with the false behavior now being the only possible behavior.
The trigger configuration included setting this property to be false: since the property no longer exists, this has been removed.


Former-commit-id: f3cb427712c2dae341a784d04843d817a76f99d2

This reverts commit cdf718bf [formerly 9d67f5b8].


Former-commit-id: 008b62d0

This reverts commit 9156e32a [formerly f73d17d3], reversing
changes made to 27d106d1 [formerly 8f2f5188].


Former-commit-id: 9d67f5b8

Veronique Boisvert authored 7 years ago and Atlas Nightlybuild committed 7 years ago

Switch all instances of TrigMuonEFTrackIsolationTool to use dz<3mm (ATR-17239)

See merge request atlas/athena!7058

(cherry picked from commit b0a55b8c [formerly 2af72ccb53058d322e9127b2c064f0b3a3b4325a])

b1066cba TrigMuonEFConfig: Switch all instances of TrackIsolationTool to use dz<3mm (was 6mm)

Former-commit-id: 8dd66b2218a9135e08b6bedaef2191935f246c6e

Remove setup of pflow items, which are never used and depend on a tool about to be deprecated (RetrievePFOTool).


Former-commit-id: 2d4c3b19d639c60b331fe15ee6ba61c748b8333c

The biggest issue was related to ToolFactory now by default producing
private tools. Initally I tried to change some trigger code to use private
tools, but that proved more difficult than I expected, so instead I changed
the configuration to make public tool instances. I changed the initialization
of EMShowerBuilder.cxx to work propery for trigger (since for trigger the
cell container is passed in and doesn't need to use a handle), and I removed
unnecessary includes to IEMBremCollectionBuilder.h in the trigger code, since
that file no longer exists.


Former-commit-id: 31d6d581ade78dca798cdc10cd41369efb791e6d