Skip to content

DerivationFrameWorkMuons - Declare data dependencies

Johannes Junggeburth requested to merge jojungge/athena:update_DerivMuons into master
    1. commit: Clang-format all C++ classes
    1. commit: Delete CellsOfMuons, ClusterNearMuon, MuonIDCovMatrixDecorator, mcpDecorator as these are never setup in any JO. Declare data-dependencies of the MuonJetDrTool, MuonTruthClassifierFallBack
        1. commit: Refactor of the MuonTruthIsolationTool. Caching of the TruthParticles into a map chopping the container into smaller containers according to their phi sector. Declaration of decorations written by the tool. Ignore neutrinos now from the isolation.
    1. commit: Refactor the IDTrackCaloDepositsDecoratorTool. It's now an Derivation::IAugmentationTool which is no longer meant to be instantiated together with the diMuonTaggingTool. The change became necessary as we must ensure that all decorations which we want to augment are written at least one time per event. The tool can either decorate the deposits of all muons or of the ID tracks associated with them. In the latter case, the first ID track in the collection is also decorated if the muon container is empty. Actually, I do not really know whether this could cause problems if this track is then thinned from the output, but I guess not. Another very important change that I made is that the decorations are now stored in a three separate vectors containing the energy loss, deposits and the depost types, respectively. Also here the figure of merit is file-integrity as well as not wasting disk-memory if for instance no HEC modules were hit. For further post processing one must schedule an algorithm that transform the new representation back to the old one (Tagging here @dcieri explicitly)
    1. commit: Refactor the IDTrackCaloDepositsDecoratorTool. The tool has been refactored analogously to the IDTrackCaloDepositsDecoratorTool. But here no output format has changed.
    1. commit: Refactor the diMuonTaggingTool. Actually I need to have a second thought about this tool as this in the version of today should be rather a thinningtool than an augmentationtool. But let's keep the current version for now.

Tagging: @mvanadia, @gabarone, @dcieri

Closes ATLASMCP-96

Edited by Johannes Junggeburth

Merge request reports