diff --git a/Reconstruction/egamma/egammaAlgs/doc/packagedoc.h b/Reconstruction/egamma/egammaAlgs/doc/packagedoc.h index 7b223a25191ebd3170d60141a82a684519f1fc7b..852636b6a806d64bbb9b6d72d44f391ba1b2a07f 100644 --- a/Reconstruction/egamma/egammaAlgs/doc/packagedoc.h +++ b/Reconstruction/egamma/egammaAlgs/doc/packagedoc.h @@ -6,60 +6,7 @@ @page egammaAlgs_page egammaAlgs Package -This package is responsible for electron and photon reconstruction and identification algorithms. - -At present, three electron and photon reconstruction and identication algorithms have been implemented in the ATLAS ofine reconstruction software, all integrated into one single package and a common event data model. - - The standard one, implemented in egammaBuilder and managed through the python script egammaGetter, which is seeded from the electromagnetic (EM) calorimeters, starts from clusters reconstructed in the calorimeters by the SlidingWindowFinder (https://twiki.cern.ch/twiki/bin/view/Atlas/SlidingWindowClustering#Sliding_Window_Clustering). For each cluster, with ET(seed)>3GeV in the cluster container the best inner detector track is searched for within a given E/p range. The complexity of the extrapolation (coordinate systems, etc) is hidden in ExtrapolateToCaloTool. It then builds identification variables -From release 13 egamma objects, having a track, which match a conversion (reconstructed by the conversion reconstruction algorithm) are considered as "Photon" and not "Electron". - - - A second algorithm, implemented in softeBuilder and managed through the python script softeGetter, which is seeded from the inner detector tracks, is optimized for electrons with energies as low as a few GeV, and selects good-quality tracks matching a relatively isolated deposition of energy in the EM calorimeters. The identication variables are then calculated in the same way as for the standard algorithm. - - - A third algorithm, implemented in egammaForwardBuilder and managed through the python script egammaForwardGetter,is dedicated to the reconstruction and identification of electrons in the forward region of ATLAS (2.5<|eta|<4.9). In contrast to the softe and egamma builders the algorithm can use only the information from the calorimeters, as the tracking system is limited to |eta|<2.5, and the topological clusters (instead of SW clusters). The pre-selection and ID are done in the same algorithm. The variables used to discriminant between electron and hadrons are defined as the topo cluster moments or combination of them. This is done separately in two eta bins: the EMEC IW and the FCal using a cut based technic. The forward electron AUTHOR is 8. - -The different algorithms are run independently during reconstruction. -There are three collections in the ESD (ElectronCollection, PhotonCollection -and softeCollection) as well as two details containers (egDetailContainer and -SofteDetailContainer), which are necessary to access the shower variables. -Therefore a true electron may be found with the standard (cluster-based) and -softe (track-based) collections. - - -At AOD level, Electrons and Photons are build by egammaAODBuilder. -They are differentiated by a loose track matching the electromagnetic cluster -and to a conversion object. -The egamma objects, having a track, -which match a conversion are considered as "Photon" and not "Electron". -The overlap for electrons reconstructed by the cluster-based and the track-based algorithms is resolved when the AODs are constructed: if the track is common -to a candidate soft-electron AND a standard egamma, then the objects are -considered to be the same. - - -There are two bits allocated for the "authorship" of each egamma object, see -egammaEvent/egammaParamDefs.h for possible authors. - -For both electrons and photons the barycentre of the electromagnetic shower is -estimated in each sampling. -The electromagnetic shower-shape variables are calculated with respect to -these barycentres in EMShowerBuilder. -Note that there are also two algorithms responsible for bremsstrahlung -recovery : EMBremsstrahlungBuilder re-calculates track pT accounting for -energy of the photons emitted in a hard bremsstrahlung. -EMTrkRefitter implements use od DNA and GSF fitters. - -EMPIDBuilder tool is used for electron and photon identification. -The PID variables can be accessed for both reconstruction algorithms : -egammaID(egammaPID::MYFAVOURITEID): -identification information (including isem). For possible values of MYFAVOURITEID see egammaEvent/egammaPIDdefs.h. -As references loose, medium and tight selections are provided. -The possible algorithms as well as discriminating variables (see for example EPiVariables for electron -vs jets discrimination variables, used by the log-likelihood and neural-network procedures), -which enter in the definition of these PID algorithms, are described in -https://twiki.cern.ch/twiki/bin/view/Atlas/ElectronReconstruction. -In addition to cut based (isem) identification, results of the following multivariate analysis tools are available: likelihood, neutral network, H-matrix and boosted decision trees. - -Other algorithms are avaliable in the egammaRec package : - -egammaTrackSlimmer : algorithm to put out the slimmed track collection +This package is responsible for electron and photon reconstruction and identification algorithms. More information about the main algorithms is provided in egammaRec. There are also algorithms here to perform GSF refitting, and to build conversion vertices. These are run as separate algorithms before the standard (or topo) egamma. An algorithm to perfomr truth associations is also provided, as well as some utility algorithms. @author H. Ma <hma@bnl.gov> @@ -70,5 +17,7 @@ egammaTrackSlimmer : algorithm to put out the slimmed track collection @author F. Derue <derue@lpnhe.in2p3.fr> @author T. Koffas <Thomas.Koffas@cern.ch> @author J. Mitrevski <Jovan.Mitrevski@cern.ch> +@author C. Anastopoulos <Christos.Anastopoulos@cern.ch> +@author D. Divalentino <david.di.valentino@cern.ch> */ diff --git a/Reconstruction/egamma/egammaAlgs/egammaAlgs/topoEgammaBuilder.h b/Reconstruction/egamma/egammaAlgs/egammaAlgs/topoEgammaBuilder.h index 2304a7d07ea84d29dc2c93a4963ab17fd9a15630..d3b5e183d95be11ada2b0755ada64e9f26325ad1 100755 --- a/Reconstruction/egamma/egammaAlgs/egammaAlgs/topoEgammaBuilder.h +++ b/Reconstruction/egamma/egammaAlgs/egammaAlgs/topoEgammaBuilder.h @@ -5,23 +5,10 @@ #ifndef EGAMMAREC_TOPOEGAMMABUILDER_H #define EGAMMAREC_TOPOEGAMMABUILDER_H /** - @class topoEgammaBuilder - top-Algorithm which creates an egammaObjectCollection. - It retrieves data objects from TDS, and calls the subalgorithms to - make the pieces in egamma classs. The subclass should inherit from - topoEgammaBuilder, and implement the method build(egamma*) - - It is seeded from the electromagnetic (EM) calorimeters and starts from clusters reconstructed in the calorimeters by the SlidingWindowFinder (https://twiki.cern.ch/twiki/bin/view/Atlas/SlidingWindowClustering#Sliding_Window_Clustering). For each cluster, with ET(seed)>3GeV in the cluster container the best inner detector track is searched for within a given E/p range. The complexity of the extrapolation (coordinate systems, etc) is hidden in ExtrapolateToCaloTool. It then builds identification variables - - The electromagnetic shower shape variables are calculated in the EMShowerBuilder tool with respect to these hot cells; - - Some variables combining inner detector and electromagnetic calorimeter information, like E/p, are build; - - In order to minimize amount of fakes, in particular in jets, pre-selection cuts are applied in egammaSelectSETool; - - Discriminating variables are build and used in the EMPIDBuilder tool based on shower shapes in the electromagnetic calorimeter and information from the inner detector; - - Possible matching to a conversion object is done through the tool EMConversionBuilder; - - Bremsstrahlung recovery information is obtained through the tools EMBremsstrahlungBuilder and EMTrkRefitter; - - the 4-momentum is build in EMFourMomBuilder. - - apply photon recovery - - apply photon post-processing to recalculate correctly - information for these recovered photons + @class topoEgammaBuilder + + The main electron and photon reconstruction algorithm, based on topoclusters. It attempts to recover electrons that emit brem photons by creating "superclusters." Superclusters can also be used in photons to especially help converted photons. This is the default egamma algorithm for central electrons and photons as of release 21. + */ // INCLUDE HEADER FILES: diff --git a/Reconstruction/egamma/egammaRec/doc/packagedoc.h b/Reconstruction/egamma/egammaRec/doc/packagedoc.h index 416779286fd476b9e69bdfd03a14f3436c7aeedf..d6f05d6e87b0e0cde47b2c56594984d84c5b7a9f 100644 --- a/Reconstruction/egamma/egammaRec/doc/packagedoc.h +++ b/Reconstruction/egamma/egammaRec/doc/packagedoc.h @@ -8,67 +8,15 @@ This package is responsible for electron and photon reconstruction and identification algorithms. -At present, three electron and photon reconstruction and identication algorithms have been implemented in the ATLAS ofine reconstruction software, all integrated into one single package and a common event data model. - - The standard one, implemented in egammaBuilder and managed through the python script egammaGetter, which is seeded from the electromagnetic (EM) calorimeters, starts from clusters reconstructed in the calorimeters by the SlidingWindowFinder (https://twiki.cern.ch/twiki/bin/view/Atlas/SlidingWindowClustering#Sliding_Window_Clustering). For each cluster, with ET(seed)>3GeV in the cluster container the best inner detector track is searched for within a given E/p range. The complexity of the extrapolation (coordinate systems, etc) is hidden in ExtrapolateToCaloTool. It then builds identification variables -From release 13 egamma objects, having a track, which match a conversion (reconstructed by the conversion reconstruction algorithm) are considered as "Photon" and not "Electron". +At present, a few electron and photon reconstruction and identication algorithms have been implemented in the ATLAS ofine reconstruction software. The configuration is handled in this package, while the main algorithms are in Reconstruction/egamma/egammaAlgs. - - A second algorithm, implemented in softeBuilder and managed through the python script softeGetter, which is seeded from the inner detector tracks, is optimized for electrons with energies as low as a few GeV, and selects good-quality tracks matching a relatively isolated deposition of energy in the EM calorimeters. The identication variables are then calculated in the same way as for the standard algorithm. + - The old one, implemented in egammaBuilder and managed through the python script egammaGetter, which is seeded from the electromagnetic (EM) calorimeters, starts from clusters reconstructed in the calorimeters by the SlidingWindowFinder (https://twiki.cern.ch/twiki/bin/view/Atlas/SlidingWindowClustering#Sliding_Window_Clustering). For each cluster, with ET(seed)>3GeV in the cluster container the best inner detector track is searched for within a given E/p range. The complexity of the extrapolation (coordinate systems, etc) is hidden in ExtrapolateToCaloTool. It then builds identification variables. As of release 21 this is no longer run. - - A third algorithm, implemented in egammaForwardBuilder and managed through the python script egammaForwardGetter,is dedicated to the reconstruction and identification of electrons in the forward region of ATLAS (2.5<|eta|<4.9). In contrast to the softe and egamma builders the algorithm can use only the information from the calorimeters, as the tracking system is limited to |eta|<2.5, and the topological clusters (instead of SW clusters). The pre-selection and ID are done in the same algorithm. The variables used to discriminant between electron and hadrons are defined as the topo cluster moments or combination of them. This is done separately in two eta bins: the EMEC IW and the FCal using a cut based technic. The forward electron AUTHOR is 8. + - An alternative algorithm, implemented by topoEgammaBuilder and also managed by the python script egammaGetter, is based on topoclusters. It attempts to recover electrons that emit brem photons by creating "superclusters." Superclusters can also be used in photons to especially help converted photons. This is the default egamma algorithm for central electrons and photons as of release 21. -The different algorithms are run independently during reconstruction. -There are three collections in the ESD (ElectronCollection, PhotonCollection -and softeCollection) as well as two details containers (egDetailContainer and -SofteDetailContainer), which are necessary to access the shower variables. -Therefore a true electron may be found with the standard (cluster-based) and -softe (track-based) collections. - - -At AOD level, Electrons and Photons are build by egammaAODBuilder. -They are differentiated by a loose track matching the electromagnetic cluster -and to a conversion object. -The egamma objects, having a track, -which match a conversion are considered as "Photon" and not "Electron". -The overlap for electrons reconstructed by the cluster-based and the track-based algorithms is resolved when the AODs are constructed: if the track is common -to a candidate soft-electron AND a standard egamma, then the objects are -considered to be the same. - - -There are two bits allocated for the "authorship" of each egamma object, see -egammaEvent/egammaParamDefs.h for possible authors. - -For both electrons and photons the barycentre of the electromagnetic shower is -estimated in each sampling. -The electromagnetic shower-shape variables are calculated with respect to -these barycentres in EMShowerBuilder. -Note that there are also two algorithms responsible for bremsstrahlung -recovery : EMBremsstrahlungBuilder re-calculates track pT accounting for -energy of the photons emitted in a hard bremsstrahlung. -EMTrkRefitter implements use od DNA and GSF fitters. - -EMPIDBuilder tool is used for electron and photon identification. -The PID variables can be accessed for both reconstruction algorithms : -egammaID(egammaPID::MYFAVOURITEID): -identification information (including isem). For possible values of MYFAVOURITEID see egammaEvent/egammaPIDdefs.h. -As references loose, medium and tight selections are provided. -The possible algorithms as well as discriminating variables (see for example EPiVariables for electron -vs jets discrimination variables, used by the log-likelihood and neural-network procedures), -which enter in the definition of these PID algorithms, are described in -https://twiki.cern.ch/twiki/bin/view/Atlas/ElectronReconstruction. -In addition to cut based (isem) identification, results of the following multivariate analysis tools are available: likelihood, neutral network, H-matrix and boosted decision trees. - -Other algorithms are avaliable in the egammaRec package : - -convAODRecVxCand : algorithm to recover photons conversions from the eelctron container - -egammaAODRender : algorithm to change the clustering and redo pid at AOD level - -egammaTrackSlimmer : algorithm to put out the slimmed track collection - -CBNTAA_egamma : fill the CBNT for reconstructed objects. - -CBNTAA_egTruth : fill the CBNT for truth particles + - An algorithm, implemented in egammaForwardBuilder and managed through the python script egammaForwardGetter, is dedicated to the reconstruction and identification of electrons in the forward region of ATLAS (2.5<|eta|<4.9). In contrast to the egamma and topoEgamma builders, this algorithm can use only the information from the calorimeters, as the tracking system is limited to |eta|<2.5, and the topological clusters (instead of SW clusters). The pre-selection and ID are done in the same algorithm. The variables used to discriminant between electron and hadrons are defined as the topo cluster moments or combination of them. This is done separately in two eta bins: the EMEC IW and the FCal using a cut based technic. +The EDM for the electrons and photons is defined in Event/xAOD/xAODEgamma. Three collections are produced in the xAOD, Electrons, Photons, and ForwardElectrons. @author H. Ma <hma@bnl.gov> @author A. Poppleton <Alan.Poppleton@cern.ch> @@ -77,5 +25,8 @@ CBNTAA_egTruth : fill the CBNT for truth particles @author A. Khodinov <khodinov@bnl.gov> @author F. Derue <derue@lpnhe.in2p3.fr> @author T. Koffas <Thomas.Koffas@cern.ch> +@author J. Mitrevski <Jovan.Mitrevski@cern.ch> +@author C. Anastopoulos <Christos.Anastopoulos@cern.ch> +@author D. Divalentino <david.di.valentino@cern.ch> */ diff --git a/Reconstruction/egamma/egammaTools/doc/packagedoc.h b/Reconstruction/egamma/egammaTools/doc/packagedoc.h index f635df9a25b52942c0f4d75414a0567a2aae770e..44e8c7079a8a038dcb66ff0d8a4fe47f27a3e853 100644 --- a/Reconstruction/egamma/egammaTools/doc/packagedoc.h +++ b/Reconstruction/egamma/egammaTools/doc/packagedoc.h @@ -6,10 +6,9 @@ @page egammaTools_page egammaTools Package -This package contains electron and photon reconstruction and identification tools. -These tools are called by main algorithms located in the egammaRec package. +This package contains electron and photon reconstruction and identification tools. These tools are called by main algorithms located in the egammaRec package. -Following tools are used by egammaBuilder for the reconstruction of isolated high-pT electrons and photons : +Following tools are used by egammaBuilder for the reconstruction of isolated high-pT electrons and photons (THE LIST IS OUT OF DATE) : egammaSwTool : creation of the cluster