ALPHA++ User`s manual ALEPH OO Physics Analysis
Contents
1. CS Termination Soe D How to find more information 20 20 21 21 22 22 23 24 24 27 29 29 30 32 34 37 39 43 43 44 46 47 1 Introduction 1 1 Introduction The ALEPH object oriented OO analysis package ALPHA is intended to sim plify programs for physics analysis by the use of C As explained in the AL PHA requirement list 3 the main ALPHA objectives are 1 Convert the ALEPH data preferably LEP2 data from the BOS bank style into persistent objects and write them to an Objectivity Db database 2 Rewrite a mini version of the ALEPH analysis package ALPHA in an object oriented computing language C based on the Objectivity database 3 Compare standard and OO performance with regard to efficient access of the data 4 Test the software engineered by the RD45 and the ANAPHE LHC project which evaluate data storage and analysis option for the LHC experiments 5 Provide some input experience for the LEP archive project 6 Give an opportunity to learn OO analysis programming and design To do so a C package driver has been implemented to loop over events just as ALPHA The possible input files are the standard POTs DSTs or MINIs EPIO files but also the Objectivity Db database hosting converted datasets Whereas the Objectivity database contains only a limited subset of the ALEPH data and Monte Carlo production and is used for2. px py pz charge leptons request code 300 options flavour flavour is a binary field It is computed as e 2 4r and describes which type of leptons are requested answer code 300 index from n to 0 Indicates the lepton index options E px py pz charge flavour flavour is equal to 1 for electrons 2 for muons and 3 for taus jets request code 400 options metric scheme ycut E metric is for Jade and 1 for Durham scheme is for EO and 1 for E E is used as is if ES 0 if E 0 the visible energy is used if 1 qelep is used answer code 400 index from n to O Indicates the jet index options cos0 px pz 0 is the cone half angle 46 C 3 Termination phase When the last event has been processed will terminate as it would in normal mode Just before exiting it will send the termination message sender ALPHA code 999 options 0 comment Done The external component can request an earlier program termination by sending the message sender external AlVisu code 999 options 0 comment Terminate As usual comments are meaningless D How to find more information 47 D How to find more information All the available information can be found on the ALPHA website http cern ch aleph proj alphapp In addition to this manual there is e the doxygen documentation e the up to date code e sample analysi
3. 3 QELEP make 6 jets using Durham PE Ycut 6 force 6 jets scheme 1 E scheme AlephCollection lt AlEflw 5 alefp EventInfo EflwPV if alefp size gt 5 f6jets alefp DurhamJet Ycut scheme 0 f6jetsP f6jets Pointers 33 34 QvecBase SortCriterium 3 sort the vector using the component 3 e of A4V sort f6jets begin f6jets end for int i 0 1 lt 6 i f6j i f6jets 5 i A4VO ABC fit info for int i 0 i 6 i PX REC il l f6j i xO PY REC il f6j il yO PZ REC il l f6j i zO REC il f6j i eO 6 jets no mass fixed ITYPP i 1 for all i NJET 6 NUP 0 ITF 1 ITEVOL 1 for int 0 i lt 7 i ITYPP i 1 F ABCFIT NJET NUP PX REC PY REC PZ REC PE REC ITF CVAL MO G0 ITYPP ITEVOL PX FIT PY FIT PZ FIT PE FIT CHI2T NDF IERR if IERR gt 1 1 recompute some variables using PX FIT PY FIT PZ FIT PE FIT save the chi2 value CHI2 CHI2T NDF 11 Interactive Mode The specific AlVisu documentation can be an interesting complement to this section All the already presented features are mainly used in the so called batch mode In this mode ALPHA runs over all the specified events and generaly fills a tree a ntuple Another possible mode in the interactive mode In interactive mode ALPHA will wait for a coded request from stdin and reply to stdout A typical use of this is to connect stdin stdout to the stdo
4. PbankH oopAllWrite joopAllWrite SimpleTokenizer ani SimpleTokenizer ankl perator Prunitrinity ysetDelimiters Peviltrinit setTerminators Poankitrinit Piokenize readBanksintoBos penContainer AlephObjyDbManager l AlephObjyDbManager Figure 5 Simplified UML diagram of the ALPHA Driver We can see the Aleph Session controling a set of AlephManagers derived from the base class AlephMan ager On the right the AlephCardReader accesses the configuration file Analysis class diagram 39 2 AlephloManager controls the IO operations It accesses the configuration file and the output root or hbook file 3 AlephExManager manages the main execution loop over runs and events and calls the UserEvent method Given the db type BOS or Objectivity Db a specific instance of AlephDbManager and AlephExManager is used Note that the managers are singletons Analysis class diagram Here is a simplified UML diagram of the ALPHA analysis part This diagram is not complete but illustrates the main structure of the object representation of data For the complete and most up to date diagrams see the HTML documentation on the web page 40 AlJet p scheme int p metric int PALet PAlJet PALet PALet STYPE PygetObjects PaddObject PclearO
5. ijet cout lt lt ijet gt QP lt lt 1 10 4 B tag The QIPBTAG ALPHA routine has been wrapped from FORTRAN It produces a set of b tagged jets This method is NOT defined in AlephCollection so the syntax is quite different You have first to instanciate the toolbox 1 01 mytoolbox Then just call pair lt AlephCollection lt AlBjet gt vector lt float gt gt result result mytoolbox ProduceBjets EventInfo 24 As usual EventInfo is the name of the AlphaBanks object The resulting pair contains e an AlephCollection of AlBjet e a vector of the probabilities for each track to come from the main vertex If the probability has not been computed by QIPBTAG the value is set to 0 The last element of this vector contains the event probability 10 5 Lepton identification 10 5 1 fundamental principles Muon identification is based on the search for a signal from the muon chambers associated to the tracking in the hadronic calorimeter Every track with enough energy is extrapolated to the HCAL where the number of active planes and the signal multiplicity are looked for Three quantities are defined Ny is the number of fired planes is the expected number along a muon path and No is the number of fired planes within the 10 allowed planes These quantities are shown on figure 2 100 Evt
6. AlGamp the Gampec object dome Sate ES 8 3 AlMCtruth the Monte Carlo truth information 8 4 AlTrack the track object 3x sam XX RS 8 5 AlThrust the thrust object Nak An ce 8 6 AlMuon the muon object 8 7 AlFlec the electron object DO eas eR ANA 8 8 Allet th jet object s sa iii iv 8 9 AlTau the tau object XX s aan RE HE Na 8 10 AlBjet the b jet object 9 Vertex objects 9 1 Vertex EINS eek OE doe Basan we S ce ARV 10 Event topology and physics routines 10 1 Missing energy mass momentum 10 2 Event topology clustering 2 nt s E Nude Wr NG SD Rd Xe TO Bitag e aude eiae A ARE PN ehe ie 10 5 Lepton identification 10 5 1 fundamental principles Re 10 5 2 ALPHA implementation of lepton identification 10 6 Kin miatical Dt 10 6 1 Description of ABCFIT 10 6 2 IN OUT variable description LOOS Examples se a SOG SH ANG NANG EC 11 Interactive Mode ALPHA driver Analysis class diagram C Full Interactive protocol description C 1 2 1
7. Riongitudinat and R5 distribution Rr Hard cuts 1 5 R2 2 lt R3 3 et 3 R5 give a high purity 99 and 60 efficiency An isolated electron is selected requiring R gt 3 In this case the efficiency is 99 An electron candidate is validated if e The track fullfills the previously described requirements e The track is a good track standard requirements on P TPC hits 40 20 e It is not already flagged as a muon The tau identification is the most difficult Only the decay products are seen so that we ll only see one of the following decays Those values are LEP2 specific and used for calibration purposes The actual values to use are set by the user see next section 26 p ALEPH EVENTS 2000 1500 1000 500r 2500 ALEPH Figure 3 Electron id quantities e 17 37 e e Vv 19 56 e h 7 0 neutrals gt 0 v 49 51 1 prong e h h h gt 0 neutrals v 15 18 3 prong hisa s ora K The two first cases are covered by the muon electron identification One or Three charged tracks will be looked for with additional HCAL activity making a mini jet with a mass close to the tau mass The Jade metric is used with gt and the E scheme A tau candidate is validated if e A Jade jet is formed with one or three charged tracks e Tracks are standard good track
8. Using those routines the same code should be able to produce ROOT or PAW files depending on the HIST card 5 1 During the UserInit routine The first step is the definition of the output contents To describe the tree ntuple a set of routines is provided No booking is needed anymore To define the variables wanted you can use a mix of the following routines void TheNtupleWriter gt AddOutput tpe obj chart name void TheNtupleWriter gt AddOutput tpe obj unsigned int sze void TheNtupleWriter gt AddScalableOutput char name tpe obj unsigned int size is any C type 5 1 1 Example 1 add an integer to the output list int i TheNtupleWriter 5Add0utput i vari There varl is the name of the variable Note that the object i in the example just tells the routines the type to be used and can be deleted later Another object will be used during the analysis so that no global object needs to be defined 5 1 2 Example 2 add several variables of the same type at once int i TheNtupleWriter gt AddOutput i 3 vari var2 var3 The second argument is just the number of declared variables 5 2 In UserEvent 9 5 1 3 Example 3 add an array of float of variable size float TheNtupleWriter AddScalableOutput array a 100 The third argument is here the MAXIMUM size of the array ROOT still needs this to allocate a global memory area Only the used part of this is written to t
9. float gt taidcuts theCardsReader gt TAIDcuts muons FilterMu muidcuts dOmax pmin minhit cosmax zOmax parameters are almup dOmax pmin minhit cosmax z0max nwirem r3cut r2cut01 r2cutil r2cut21 r2cuth r5cutl r5cuth elecs FilterEl almup elidcuts tausP FilterTau taidcuts pmin dOmax minhit cosmax zOmax Parameters are taken from the driver as in the example and reflect the MUID TAID and ELID parameter cards or may be set manually by the user If a card is not provided default values are used A vector of muons which can be empty must be given to the electron filtering algorithm It is used to grant that no electron is alreday flagged as a muon If an empty vector is given the test has to be done afterwards by the user Default parameters are the same as in ALPHA 10 6 Kinematical fit 29 dOmax 0 5 2 for 7 minimal distance to the beam Pmin 1 minimal lepton momentum Minhit 4 minimal number of hits in the TPC Cosmax 0 95 maximal cosine of the track beam angle ZOmax 10 minimal distance to the interaction point along z Nwirem 40 minimal number of usefull wires for dE dx R3cut 1000 upper limit on R3 dE dx R2cut01 3 lower limit for R2 in a clean HCAL area R2cut11 7 lower limit for R2 in a crack R2cut21 5 lower limit for R2 in a overlap R2cuth upper limit for R2 R5cutl 0 5 lower limit for R5 R5cuth upper limit for R5 10 6 Kinematical fit Constrained kinematical fits on an
10. org index html http cern ch aleph proj alphapp doc goals html 4 http wwwinfo cern ch asd lhc HTL 5 http root cern ch root RootDoc html 6 ALEPH Collaboration ALEPH A detector for electron positron annihilation at LEP Nucl Inst Meth A294 1990 121 7 http alephwww cern ch hansenjo ALEPH ONLY abcfit abcfit html A ALPHA driver The main class is the AlephSession Three methods are called one after the other Initial izeQ Run and Terminate Those three states are propagated to the different man agers 1 AlephDbManager controls the database access BOS or Objectivity Db and maps it to the AlphaBank 38 AlephManager Yinitialize terminate AlephManager AlephManagerSetup AlephManagerShutDown AlephSession I PalephSession Bp debugLevel int ye A MC etStatus 2 5 _theSession _ 1 sessionType 3 fon ON th AlephDbManager und theAlephloManager AlephStatus ferminate v AlephManager theAlephExManager PstartTimer stopTimer setDebugLevel lebugLevel lephSession AlephS
11. test and demo purposes all the EDIR files can be read and analysed so that a full analysis can now be performed with ALPHA The program structure is very simple see appendix A Three C methods are supplied by the user job initialisation event processing and job termination In addition a full class tree makes the analysis very easy using objects like tracks energy flows jets etc appendix B Alpha is still used internally to access data so that the standard ALPHA cards when meaningful can be used The main ALPHA physics routines are also imple mented in C to use modify and create the object classes This document describes all features of framework The main features are directly related to the ALPHA ones so that we will reg ularly refer to the ALPHA manual Whenever possible the ALPHA conventions have been used so that ALPHA should be natural to use for an ALPHA user In addition the outline of this manual is greatly inspired from the ALPHA one 1 Finaly the present manual should be complemented by the consultation of the website http cern ch aleph proj alphapp It contains all the infor mation needed to start with ALPHA and all the documented code processed with Doxygen 2 2 Getting started For a fast start visit http cern ch aleph proj alphapp doc start html The platform supported by ALPHA is Linux in particular the Ixplus environ ment To be able
12. 75 E muons ALEPH sol pions a 25 0 4 EN olf 6 ae 0 0 2 0 4 0 6 0 8 1 12 Ns Evt 100 75 50 b 25 0 2 4 6 8 10 102 951 a Ld fire 1 1 1 1 1 1 1 1 2 4 16 18 2 22 24 Figure 2 Muon id quantities A muon candidate is validated if e The track is a good track standard requirements on P TPC hits dO 20 10 5 Lepton identification 25 e Nyir Nexp gt 04 Nio gt 4 and multiplicity lt 1 4 e There is a matching muon chamber signal Electron identification is a bit more tricky There are two possible ways to iden tify an electron the energy pattern in the ECAL or the TPC dE dx information might be used The first possibility is more powerful and will be prefered as long as the electron is not in a crack or an overlap A set of estimators is built at re construction time which describes in units of standard deviation the difference between the expected mean value and the actual value e energetic balance E4 P E4 P 02 e R2 transverse shape R3 longitudinal shape 03 R4 longitudinal shape 24 04 R5 specific ionisation Ll E and p are the energy and the momentum E is the energy deposit in the 4 closest ECAL towers The parameters a and 5 describe the longitudinal profile using the following fit Le ghehe 79 95e Figure shows the R2 distribution Riransverse distribution
13. ALEPH Collaboration CERN ALPHA User s manual ALEPH OO Physics Analysis Package release 3 4 R Cavanaugh C Delaere G Dissertori K Huettmann V Lemaitre O van der Aa CONTENTS Contents 1 6 7 8 Introduction Getting started Input cards User routines Creating histograms and ntuples 5 1 During the UserlInit routine 5 1 1 Example 1 add an integer to the outputlist 5 1 2 Example 2 add several variables of the same type at once 5 1 3 Example 3 add an array of float of variable size o lsb SOME COMERS s ec w crea o eais en DA c MWUSELEVENG u oso e we ale EN OO UA B 5 2 1 Record an instant picture of a single variable 5 2 2 Record an instant picture of a scalable variable array 5 3 At the end of UserEvent 5 4 More about the HBOOK interface 5 5 More about Event and Run informations 61 AlEvent methods 6 2 AlRun methods Handling objects 7 1 Access to the objects 72 The RIODIECECIASS S Gate a deux 7 3 The QvecBase lt 7 4 The QvrtBase class Tracks and vectorial objects 8 1 AlEflw the Enery Flow object sea nonr ee We 8 2
14. AlTrack AIMCtruth AlGamp Bp int AITrack AIMCtruth PAIEfiw AITrack AIMCtruth AIGamp AlTrack AIMCtruth BsetEfType NVO PsetMotherVector PyetEfType PsetDaughterVector Feci NE jetTrack jetDaughterVector Mock PNR Punlock MR 11 PND 11 P00 DIM 24 91120 lt 9 5 0 4 lt i TMmean PTL2sum i NSsum ADmean PsetMatchingVector AgetMatchingVector Figure 6 Simplified UML diagram for ALPHA Part 1 Analysis class diagram QVEC jext 7 QEXT float 10 ECHT int NO int LCHT int eN int int float FCOT int int LCOT int Sx fost NCO float FIST int QY float Fi LIST int 2 float NIST int float FAST int ga ga LAST int float NAST int float FRET int 5 LRET int gKS int NRET int int FEFT int 2 QDET LEFT int Vv jung 0 1 NEFT int int m LNET int NNET int eBM in 14 N FGAT int eK int Y LGAT int float Te NGAT int 2 float NI NG SD float w LJET int 52 float NJET int float FMCT int
15. US float 18 LMCT int int PAssign N FVOT int 1 LVOT int NVOT int FDCT int LDCT int int int LREV int NREV int Seamer y AlphaBanks GampPV SERVO PEfwPV PTrackV j TrackPV ih MCtruthV MCtruthPV MuonV lecV Heavily ElecPV modified ObjectPV PEvent PRun PaddObject 4faddObject Figure 7 Simplified UML diagram for ALPHA Part 2 0 0 1 qhac qhac KTA 9 Heavily AlephColleetion modified AlRun Spaext QEXT AIRun AIRun Eneroy ExperimentNumber number PL epHeader Run sETqhacl SETgext lt PAlephCollection Ng ereen rint N Pprint2 N looperase Pointers AThrust PSphericity Paplanarity Pplanarity DurhamJet JadeJet ATauJn DurhamYn FilterMu rilterEI FilterTau BGetSum GetMiss L J JadeProtoJet Scheme int AlEvent Slsetup int mask1 int mask int mask3 int tmask int AlEvent AlEvent Time number e12 xiumok isTrigger GetRawTrigger Header GeneratorSta
16. Vertex Note that the vertex is passed as a pointer The constraints can be changed anytime e g in order to test different mass hypotheses The actual fit is performed by calling the method doFit with an AlephCollection of pointers to QvecBase objects AlTrack AlUserTrack or AlEflow The number of input tracks in the collection must not be larger than 10 AlephCollection lt QvecBase gt my collection 10 Event topology and physics routines 21 track selection AlUserTrack mother track AlUsrVertex end vertex bool failed my vfit doFit my selection mother track end vertex If the fit has been sucessful the return value will be false and the reconstructed track and the vertex it is pointing to will be returned through the AlUsrVertex and AlUserTrack argument respectively The AlUserTrack will also have a pointer to the end vertex and a vector of pointers to the tracks it has been build of The x of the fit is accessible through the QVCHIF method of AlUserVertex Kinematic quantities and covariance matrices can likewise be accessed by the respective methods of AlUsrVertex and AlUserTrack Note that the track parameters of the input tracks remain unchanged and that no new objects are added to the ALPHA banks 10 Event topology and physics routines the routines described in this chapter perform loops over the elements of an Ale phCollection The objects to be considered are selected by removing bad objects from t
17. an do there all what you are used to do in UserEvent In addition there is a method void SendMessage int code vector lt float gt amp options string comment available that can be used to send well formated messages with code options and com ment See annexe C for more details Nevertheless you must respect some conventions in order to have AlVisu understanding your function When returning the objects resulting from your calculation the first option called the index must be a decreaing number equal to zero for the last object returned The second argument must be the object type 1 a track 2 an energy flow 3 a lepton 4 a jet The next options are then function of the object type e for a track px py pz ch e for an eflow E px py pz ch e for a lepton E px py pz ch flavour 1 e 2 4 3 7 e for a jet cos px py pz cos 0 gives the cone half angle An example can be found in the test1 cpp file Here it is 36 class testclass public AlephAbstractInteractiveFunction public testclass AlephInteractiveHandler ptr AlephAbstractInteractiveFunction ptr virtual string Name return test virtual int Code return 555 virtual void Run vector lt float gt amp options AlphaBanks amp EventInfo 1 dummy routine returns always the same track eflow and the same jet vector lt float gt output output push_back 2 output push_back 1 a track output push_back 3 px output p
18. bjects PsetScheme PygetScheme PsetMetric PygetMetric lt 02 PLock PunLock AlTau p nch int y ech float ParTau ParTau ParTau BsetNch BgetNch setEch PgetEch PsetEfw DgetEfiw AlThrust AlMuon AlElec AIThrust AlMuon AIElec AIThrust AlMuon AlElec STYPE BgetThrustDirection PLock Lock BgetThrustValue SunLock BunLock 1 SsetEfiw setTrack setTrack SyetTrack BF Fo 1 sRO DRI H DM R2 Bs R3 SINO R4 BR5 DRI BEC BP 1 E2 E30 BER AlObject QVEC float 10 7 NO int L N int float int float QvecBase ta ma 402 float float s float LA P float pa float Refl C int Modified gt es int lock 0 1 gCL int eQD QDET K OV int 0 EV int ND zint Weel ink NM Int BM int na QvecLink pm BsetPointer 4 BQvec arp lod 4 gt Pose 457 float ODBS20 eco mont Q7BS2 US float 18 MU PAssign 1 AlEfw
19. c In the option PARID i kmn n chooses the kind of jet parametrisation n 0 ALEPH like parametrisation gt HN gt E pi r ai pim bi u theta ci u phi 1 DELPHI like parametrisation gt a n gt sees exp ai pim bi u theta ci u phi B DELPHI parametrisation has 2 main advantages the ai distribution looks like more Gaussian for low momentum jet and it avoids fit results with reverse jet mo mentum The option ITEVOL kmn fixes the choices on the variation of ai bi ci parameters with jet momentum and angle on the covariance matrix and on the reference system use the particles true or measured values for binning The pre ferred values are PARID i 1 and ITEVOL 1 for the fit which then need the file aibi evol 1111 001 dat in case NJET 4 or aibi evol 11111 001 dat in case NJET 5 Examples of different parametrisation files used by ALEPH WW analysis at 183 GeV and 189 GeV can be found in 7 The description of the IN OUT variables of the routine is presented bellow fol lowed by an example of how to use ABCFIT in ALPHA 10 6 2 IN OUT variable description F ABCFIT NJET NUP PX REC PY REC PZ REC PE REC ITF CVAL MO GO PARID ITEVOL PX FIT PY FIT PZ FIT PE FIT CHI2T NDF IERR IN NJET Number of objects in event lt 7 NUP Number of Unmeasured Particles in event must be last in p _rec PX REC 7 x momentum of up to 7 objects PY REC 7 y mo
20. collection begin itrack lt mycollection endO itrack cout lt lt itrack gt QX lt lt endl or for unsigned int i 0 i lt mycollection size i cout lt lt mycollectionlil gt QX lt lt endl Although the second solution is shorter in terms of typing and seem more standard it is in fact far more time consuming To erase elements in an AlephCollection the looperase method is usefull 7 2 The AlObject class 13 for AlephCollection lt AlTrack gt iterator itrack mycollection begin itrack lt mycollection end itrack if idx gt QP lt 1 looperase itrack Finally two collections be merged by append AlephCollection lt type gt and you can create a vector of pointers to the objects in the vector with the Point ers method AlephCollection lt AlJet gt myjets AlephCollection lt AlJet gt mypointers myjets Pointers 7 2 The AlObject class All the ALPHA objects are derived from AlObject AlObject is a pure virtual class definig thus the general interface e HepLorentzVector A4V returns a Lorentz vector e ALEPHTYPE TYPE returns the object type e float returns the charge e more The two classes directly deriving from AlObject are QvecBase for all track like objects and QvrtBase for all vertices 7 3 The QvecBase class The methods of QvecBase define some standard ALPHA routines common to all track l
21. cts fitted energy of up to NPARTICLES objects total chi2 lt number of degree of freedom error flag results are irrelevant gt 0 results fulfils constrains and be used 10 6 3 Example In this piece of ALPHA code one forces the event to be clustered into 6 jets and kinematical fit is performed with the option ITF 1 i e momentum and energy conserva tion The options chosen ITEVOL 1 diagonal reco binning and PARID i 1 fit parametrisation a la DELPHI and the code needs the parametrisation file available on the for details on IERR ouput value see http alephwww cern ch hansenjo ALEPH ONLY abcfit callstruc html web aibi_evol_111111_001 dat define input and output of ABCFIT float PX REC 7 float PY REC T 10 6 Kinematical fit float PZ REC T float REC T float PX FIT 71 float PY FIT T float PZ FIT 71 float FIT 7 float CVAL 4 float 3 float G0 3 float CHI2T float CHI2 int NJET int NUP int ITF int ITEVOL int ITYPP 7 int NDF int IERR for int i 0 i lt 7 i PX_REC i 0 PY_REC i 0 PZ_REC i 0 PE_REC i 0 set values of mass and width for int 1 0 1 lt 3 i 1 0 for int 1 0 i3 i GO i 0 set values of four momentum constrains for the fitted system QELEP is the centre of mass energy of the initial state CVAL 0 0 CVAL 1 0 CVAL 2 0 CVAL
22. dex option index function opcode default option value comment option name 3 When all the functions are send ALPHA will send the Ready message and wait for the first event request code options comment C 2 Normal Data exchange phase During the normal phase of data exchange various messages are emited and received There are the various user s messages and some standard messages The format of the user s messages will not be described here again Please refer to the correspond section in the present manual The standard messages correspond to default functions implemented next event request code 100 options all events 1 only events with UserEvent returning true answer A set of 6 messages is send with each time the code 100 The index varies from 1 to 6 Other options are index 1 event number index 2 run number index 3 year month day hour min sec index 4 edir class index 5 msum esum index 6 acoplanarity acolinearity thrust thrust_x thrust_y thrust_z aplanarity planarity sphericity tracks request C 2 Normal Data exchange phase 45 code 200 options eflow Energy cut cos cut sanity Eflow can take 3 values 0 no request 1 tracks 2 energy flows Sanity is a pure boolean If it is 1 only good tracks eflows are selected answer code 200 index from n to O Indicates the track index options E
23. e Lorentz vector associated with the muons By defnition the Lorentz vector of a lepton is the one from the associated track 8 7 AlElec the electron object 19 8 7 AlFlec the electron object AlEflw getEflw returns the associated energy flow object AlTrack getTrack returns the associated track Note that these values should be checked for being different from NULL such as all the pointers returned by ALPHA methods In addition the class gives access to the EIDT bank Note about the Lorentz vector associated with the electrons By defnition the Lorentz vector of a lepton is the one from the associated track 8 8 AlJet the jet object Jet objects are returned by the various jet algorithms as described later in this manual list of the objects in the jet AlephCollection lt AlObject gt amp getObjects int getScheme scheme used to build the jet 0 for EO 1 for E int getMetric metric used for the jet 1 for Jade 0 for Durham 8 9 AlTau the tau object The tau object is derived from AlJet and has thus all the jet properties and methods In addition int getNch returns the number of charged tracks float getEch returns the total energy of charged tracks AlEflw getEflw returns the associated main Energy flow object By definition the main Energy flow object is the most energetic one with the same sign as the tau 8 10 AlBjet the b jet object AlBjet is a jet and thus has all the jet prope
24. er or batch Batch mode is used for most of the analysis jobs In interactive mode the software will wait after each event for a user request generaly generated by an external software Up to now only AlVisu the OO visualisation software is using this feature name of the file including the FILI statements for obejectivity input The list of FILI statements included in this file is appended to the any list directly included in the file ALPHA FILI ALPHA FILO ALPHA READ ALPHA SRUN 3 Input cards 5 HIST SEED ACAR MUID ELID TAID USER the output histogram file Depending on the file extension the output will be produced by ROOT root or HBOOK everything else random generator seed ALPHA card complement When reading EDIR you can specify additional ALPHA cards in a separate file A standard ALPHA cardsfile is built using those cards the info specified in the ALPHA input cardfile and some additional cards are addede automatically EFLW BSIZ 0 0120 0 0007 QFND DWIN 1 0 035 NSEQ ENDQ cuts for muon identification see the section 10 5 2 cuts for electron identification see the section 10 5 2 cuts for tau identification see the section 10 5 2 The user can define his own cards This card allows to define quantities that you will use in your own code cali brations cuts debug various options The synthax is the following USER name valuel value2 The first argument is the tag y
25. ession AlephDbManager AlephExManager AlephloManager currentRun long STheAlephDbManager currentEvt long sTheAlephloManager Biniialize theAlephSession sTheAlephExManager un ransactionType Fan terminate etTransactionType serlnit heCardsReader 0 1 startTransaction un lephloManager 1 topTransaction PUserEvent i AlephloManager Paaratandhokirarsactong Perminate 041 penExistingDb penNewDb TA etCurrentRunEvt IbName maa PoetCurrentRun PsetDbName XD MetCurrentEvt a JAlephDbManager 4 Mhesession lt i AlephDbManager heDbm heObjyDbm helom AlephCardsReader MneCardsReader Gp Net ong from lt unspecified gt lephExManager ES PoetNevt 0 1 AlephCardsReader BoetClass T AgetDbFiles Rospo Typo jetTaType oetAlphaFilo AlephObjyDbManager PE jetAlphaRead Piritialize LN etHist orina 1 MUIDcuts startTransaction PELIDcuts stopTransaction TAlDcuts commitAndHoldTransaction g orepareAlphaCards PopenExistingDb AlephEpioExManager AlephObjyExManager penNewDb HepRef initialize Minitialize ung un runContH Perminate Merminate eviContH lephEpioExManager lephObjyExManager loopAll runH AlephEpioExManager AlephObjyExManager loopAllRead SimpleTokenizer
26. event be performed in thanks to an interface to the fortran ABCFIT program A detailed description of the ABCFIT program can be found 7 The original code used ALPHA was taken in the directory shift50 aleph wwtf Common_Ntuples tools w1 31 abcfit and corresponds to a slightly upgraded version compared to one described in 7 10 6 1 Description of ABCFIT The ABCFIT package is designed to do constrained kinematic fitting of multi particle events The ABCFIT package employs an analytical approach by means of Newton Raphson minimisation combined with Lagrange multipliers resulting in an iterative algorithm Some key features include variable number of particles maximum 7 are allowed the concept of unmeasured particles numerous parameterisations are included and it is possible to introduce addi tional non diagonal input covariance matrices of fit parameters parameter values and covariance matrices can be determined as a function of true or measured quantities user specified constrain values Gaussian or Breit Wigner behaviour of masses 30 To run the program an external files is needed which name depends on the tions chosen for the fit the file aibi evol itypp njet itevol kOn dat is used for parametrisation definition and parameter evolution Each particle mo mentum is described via a longitudinal momentum scale a and two transverse components b and
27. hat you don t switch of Px Py and Pz constraints by accident value of up to 3 mass constrains e g mw mw width of up to 3 mass constrains e g w width kmn Fit parameterisation a la ALEPH Fit parameterisation a la DELPHI Fit parameterisation using P theta phi Fit parameterisation using Px Py Pz Fitted jet energies scale with fitted momenta Fitted jet energies is determined from fixed input jet mass and fitted momenta 2 Fit jet mass using a 4 th parameter a la ALEPH M_fit d M_reco OWN H N ITEVOL QUT FIT 7 PY FIT 7 PZ_FIT 7 PE_FIT 7 CHI2T NDF IERR IERR X KF X Xo XX X KF KF ERROR OUTPUT 3 Fit jet mass using a 4 th parameter la DELPHI M fit exp d M reco 4 Fit jet mass using a 4 th parameter E fit exp d log E reco P reco 1 1 P fit 0 Full correction on jet momentum and T 1 Only correction on jet momentum 2 Full correction on jet momentum but longitudinal parameter is assumed to be unmeasured 3 switch on k 1 _ _ k 2 kmn 0 Evolution with flavour jet characteristics 0 Diagonal input covariance matrix 1 Full non diagonal input covariance matrix 0 binning in reco values for correction factors 1 binning in true values for correction factors fitted X momentum of up to NPARTICLES objects fitted Y momentum of up to NPARTICLES objects fitted Z momentum of up to NPARTICLES obje
28. he ROOT file 5 1 4 Some coments e If the class object is not a basic type it will only work with ROOT cfr the ROOT documentation for more details e Up to now the ScalableOutput is not implemented for HBOOK It is intended to interface the column wise ntuples 5 2 In UserEvent In the UserEvent routine the specific event is analysed and the reconstructed infor mation is stored in the specific ntuple tree variables 5 2 1 Record an instant picture of a single variable void TheNtupleWriter gt Keep char name entry value example int myvar 3 TheNtupleWriter gt Keep output myvar 5 2 2 Record an instant picture of a scalable variable array void TheNtupleWriter gt Keep char name entry startofarray int size example float myarray 10 0 1 2 3 4 5 6 7 8 9 TheNtupleWriter gt 10 There also specific methods to store STL vectors HepVectors etc In each case the name must correspond to a Scalable Output void TheNtupleWriter 5KeepV chart name vector lt entry gt value void TheNtupleWriter gt Keep char name HepLorentzVector amp value void TheNtupleWriter gt Keep char name Hep3Vector amp value The object will be stored as a float array 10 5 3 At the end of UserEvent At the end of the routine you just call TheNtupleWriter gt Fill to store everything in the histograming file if the event is selected e g after
29. he collection In addition locked objects are not used 10 1 Missing energy mass momentum Two methods can be used to compute such quantities e GetSum computes the sum QvecBase Object e GetMiss qelep computes the missing QvecBase object with respect to qelep In both cases the result is returned as a full QvecBase object having a Lorentz vector In the first case this vector is the sum of all the vectors of the collection the method is applied on In the second case this is the result of the difference of this sum with a vector of energy qelep in most of the cases the LEP energy and momentum O From the resulting objects it is easy to get the enery QEO mass or momentum Example To compute the missing transverse momentum 22 get the collection of energy flow objects AlephCollection lt AlEflw gt myobjects EventInfo EflwPV compute the mmissing object QvecBase mymissingobject myobjects GetMiss EventInfo Event Energy get transverse momentum cout lt lt mymissingobject QPT lt lt endl 10 2 Event topology They are four general topology routines e float Sphericity e float Aplanarity e float Planarity e AlThrust AThrust The AlThrust object includes the thrust direction and value They are accessed using the getThrustDirection and getThrustValue methods Example To compute the thrust get the collection of energy flow objects AlephCollectio
30. he protocol are 1 It s based on the standard input output The advantages are that it can be directly used by the user and that it can be piped for example with ssh 2 Messages are simple mostly human readable strings 3 Build on a request reply basis The external component will always request some data using a numeric code ALPHA will the reply with a set of messages contain ing the same code The only exections are the initialisation and termination phases Any message must be formated like this separator AlVisu gt lt 400 gt lt 1 gt lt 1 gt lt 3 gt lt 1 gt lt jets sender code options comment Figure 8 Standard message It is made of a variable set of substrings separated by gt lt The first string is the sender name Up to now only ALPHA and AlVisu are allowed There is then the operation code opcode followed by a set of options The first option is generaly called the index The message is terminated by a comment string generaly meaningless C 1 Initialisation phase During the initialisation ALPHA will send a list of the available functions standard and user s ones with a list of the options with default values The following sequence will be repeated for each function 1 First the function header is send code 1 options function index number of options function opcode comment function name 44 2 Then for each option code 1 options function in
31. ike objects e HepLorentzVector A4V returns a Lorentz vector e ALEPHTYPE TYPE returns the object type float QYO 020 QEO QCTO QCHO QBETAO QGAMMA The meaning of those routines is the same as for AL PHA 14 Some additional methods take another QvecBase descending object as argu ment Also the ALPHA manual should be consulted for more details e float QMSQ2 QvecBase e QM2 QvecBase j e QDMSQ QvecBase j e QPPAR QvecBase j e QPPER QvecBase j e QDOT3 QvecBase j e ODOT4 QvecBase j e QCOSA QvecBase j e QDECA2 QvecBase e QDECAN QvecBase Finally ALPHA specific methods are the lock related methods and the ability to sort objects There are 3 lock related methods e Lock will lock the object setting a flag A locked object will not be used by the various algorithms The method takes as argument an optional parameter Lock 1 will lock the object recursively For example locking a lepton will lock the associated track Eflow object e In the same way unlock will unlock an object and unlock 1 will unlock recursively e isLocked returns a boolean corresponding to the Lock status true locked The QvecBase derived objects can be sorted simply using the operators The sort criterium can be selected by defining the global variable QvecBase SortCriterium The defined values are 0 1 2 3 4 respectively for E px PZ pt 7 4 The Q
32. ility Pointer to the PGAC parent giving a fake photon AlGamp ParentGivingFake Flag for fake determination int FakeEquality pointeur to PECO bank int Peco0bject 8 3 AlMCtruth the Monte Carlo truth information collection of mothers AlephCollection lt AlMCtruth gt getMotherVector collection of daughters AlephCollection lt AlMCtruth gt getDaughterVector int PA particle code int NO number of mothers int ND number of daughters char name particle name 18 8 4 AlTrack the track object It gives a direct access to the FRTL FRFT and TEXS bank entrie TLO TL20 In addition the following methods are defined float TMean mean TM float TL2sum sum of TL2 float ADmean mean of AD int NSsum total number of segments matching MC particles AlephCollection lt AlMCtruth gt amp getMatchingVector AlEflw getEflw associated energy flow 8 5 AlThrust the thrust object The thrust object is build by the thrust algorithm Hep3vector getThrustDirection returns the thrust direction float getThrustValue returns the thrust value 8 6 AlMuon the muon object AlEflw getEflw returns the associated energy flow object AlTrack getTrack returns the associated track Note that these values should be checked for being different from NULL such as all the pointers returned by methods In addition the class gives access to the MUID bank Note about th
33. mentum of up to 7 objects PZ_REC 7 z momentum of up to 7 objects PE REC 7 energy of up to 7 objects ITF mn 10 6 Kinematical fit 31 gt gt CVAL 4 3 G0 3 PARID 7 1 Four momentum and energy conservation Som j 1 njet p_fit j cval 2 Energy momentum conservation m 1 2 0 1 GO is not used 3 Energy momentum conservation alpha 1 2 m 1 2 m0 1 sigma alpha GO 1 m0 1 4 Energy momentum conservation m 1 2 m 3 4 0 5 Energy momentum conservation alpha 1 2 m 1 2 alpha 3 4 m 3 4 0 sigma alpha 1 2 GO 1 m0 1 sigma alpha 3 4 G0 2 m0 2 6 Energy momentum conservation m 1 2 m0 1 m 3 4 m0 2 7 Energy momentum conservation alpha 1 2 m 1 2 m0 1 alpha 3 4 m 3 4 m0 2 8 Energy momentum conservation E 1 2 E 3 4 equal energy In case of mass constraint Gaussian fit is used 1 In case of mass constraint Breit Wigner fit is used i e alpha parameters have a BW distribution value of 4 momentum constrains Px Py Pz E e g 0 0 0 ecm Can be used to switch of some of the constraints If abs cval 1 gt abs cval 4 Px constraint is not applied Same holds for cval 2 Py and cval 3 Pz If cval 4 1t 0 0 E constraint is not applied BUT remember to set value such t
34. n lt AlEflw gt myobjects EventInfo EflwPV compute the thrust object AlThrust mythrust myobjects AThrust get the thrust value cout lt lt mythrust getThrustValue lt lt endl 10 3 Jet clustering The content of a collection can be clustered to form jets Two kinds of metric are available Durham and Jade Two schemes can be used for each metric E energy momentum conservation and EO momentum is rescaled to get a massless particle To use the Durham metric call 10 4 B tag 23 DurhamJet float theycut int scheme float energy The scheme is 0 for and 1 for E The third argument is the event energy If it is set to O the total visible energy of the set of objects is used In the same way JadeJet can be choosen to use the Jade metric In both cases a fixed number of jets can be obtained setting the ycut to n where n is the number of jets wanted The object produced is an AlephCollection of AlJet The AlJet object is derived from QvecBase and is described in a previous section Example to compute jets get the collection of energy flow objects AlephCollection lt AlEflw gt myobjects EventInfo EflwPV call the method DurhamJet float Ycut 2 force 2 jets int scheme 1 E scheme AlephCollection lt AlJet gt Jn myobjects DurhamJet Ycut scheme 0 get the momentum of each jet for AlephCollection lt AlJet gt iterator ijet Jn begin ijet lt Jn end
35. nd fill ntuples To produce a PAW ntuple just remove the root extension in the HIST card placed in ALPHA input Note that the PAW output is unavailable for acoplanar the default compiled code due to the use of columnwise like commands still not implemented for the ALPHA HBOOK wrapper A sample script to run a job on LSF is available on the web page 4 3 Input cards The file defined by the environment variable SALPHACARDS contains the configu ration cards used by ALPHA In this file are defined the events to be processed and some options The cards available follow similar rules compared to ALPHA Comments are preceeded by a ora p Availables cards are NEVT CLAS DBTY TATY SETY READ AFII AFIO AREA ASRU event selection by number As with ALPHA there are 2 synthaxes e NEVT n read the n first events e NEVT n1 n2 read events from n1 to n2 Only events of the relevant class are read There must be only one NEVT card select the EDIR classe s to read CLASS c1 c2 c3 read events of class c1 c2 and c3 database type 3 obtions are availables e DBTY objy access the federated Objectivity Db database e DBTY epio read the standard ALPHA files POT MINI DST e DBTY epio objy read standard ALPHA files and open the Objectivity Db database for HTL 4 output transaction type read or write Up to now only read transactions are imple mented session type int
36. ou assign to the card the other are tag quan tities being passed to the code Even if any string can be used as tag it is recommended to use 4 uppercase letters like the other ALPHA cards In the user code the info will be available by theCardsReader gt getUserVards returning a vector lt pair lt string vector lt double gt gt gt For each entry in the vector the first element of the pair is the card name and the second is the vector of values 6 4 User routines Only three routines have to be provided initialisation void AlephExManager UserInit event analysis bool AlephExManager UserEvent AlphaBanks amp Event Info termination void AlephExManager UserTerm Each example code provided gives an example of those three routines In order to get the code compiled you must include the following lines at the beginning of the code include AlephExManager h include AlephCollection h In addition if you want to write in the output text file used by the driver include the fout stream as follows include lt fstream h gt extern ofstream fout Other possible includes are e include lt vector gt to directly use STL vectors e include AlToolBox h to use the Tool Box see the section 10 4 If you want your own classes to be known by the driver you have to put the declaration in UserClasses h UserClasses h is included in the driver and it should be possible to use them with ROOT see the ROOT man
37. rties and methods In addition float getBprobability returns the QIPBTAG probability for that jet 20 9 Vertex objects There are 2 vertex classes e AlVertex describes the vertices from the BOS e AlUsrVertex describes a user defined vertex The methods are the QvrtBase ones Nevertheless AlUsrVertex returns a dummy ouput for KVN QVCHIF and QVEM For a AlUsrVertex setA4V HepLorentzVector has to be used to set the position and energy of the vertex By definition the user vertex type is 6 9 1 Vertex Fitting The AlVertexFitter class provides a general interface to vertex fitting routines In the present implementation the fit algorithms are those of the Fortran YTOP package in ALEPHLIB The vertex fit refits the track parameters of a set of input tracks and calculates the 4 vector and error matrix of the new track the reconstructed decaying object at the fit vertex Depending on the fit constraints to be set three different classes are available which are derived form the general vertex fitter class An object of one of those classes must be instanciated once Vfit my_vfit simple vertex fit no additional constraints MassCVfit my_mcvfit mass constrained fit VertexCVfit my_vcvfit vertex constrained fit For the latter two additional constraints have to be specified e g double KOs_mass 0 511 my_mcvfit setMassC KOs_mass AlVertex theMainVertex event MainVertexP my vcfit setVertexC theMain
38. s 6 The total charged energy is large enough as fixed the pmin cut 10 5 Lepton identification 27 10 5 2 implementation of lepton identification To identify and manipulate leptons three classes are defined AlMuon AlElec and AlTau Classes describing electrons and muons give access to data used to select them There is also a pointer to the associated track and energy flow object The tau object is derived from AlJet It is possible to separately select electrons muons and taus In both cases the procedure is the same figure 4 Electron Muon Tau Get Energy flows from AlphaBank AlephVector lt AlEflw gt eflws bank eflwPV Get candidates from AlphaBank Get candidates from AlphaBank Gat candidates Eflows AlephVector lt AlElec gt elecs bank ElecPV AlephVector lt AlMuon gt AlephVector lt AlTau gt muons bank muonPV taus eflws AtauJn Ycut scheme metric Evis Build Pointers AlephVector lt AlTau gt tausP taus Pointers Filter the candidates Filter the candidates elecs FilterEl dOmax minhit cosmax z0max muons FilterMu d0max minhit cosmax zOmax Filter the candidates elecs FilterEl d0max minhit cosmax zOmax Figure 4 Selection procedure for electrons muons and taus Ge
39. s codes e FAQ to come In addition it is recommended to consult the ALPHA manual
40. s constructed from the standard ALPHA date and time strings Information is obtained via public data members msec sec min hour day month year In addition the comparison operators are defined int number event number int type event type int EdirClass edir class pattern int ErrorStatus detector error status float Energy event Energy QELEP float gen 12 MC e12 generator result for the event This value is generated when needed Two subsequent calls will give differents results If the seed is fixed two executions of the code will give the same result bool IsTrigger bool MINI simulate the trigger word as the alpha routine DEC TRG do The argument must be TRUE when running on MINI FALSE otherwise int GetRawTrigger returns the raw trigger word 6 2 AlRun methods int energy returns the mean run energy as stored in the LEP header bank int ExperimentNumber returns the experiment number int number returns the run number int type returns the run type 7 Handling objects 7 1 Access to the objects Objects are available as vectors filled by the AlphaBanks class just before the user routine is called Those transient objects are used during the analysis in replace ment of the persistent objects stored in the epio files or Objectivity Db databases 12 The AlephCollection is a specific class defined for this purpose The physics routines are methods of AlephCollection so that the
41. some preselection cuts 5 4 More about the HBOOK interface The methods presented in the previous section are intended to make the histogram ming easy and independant of the output format processing tool Nevertheless the capabilities of both HBOOK and ROOT are more developed and you may want to use more specific methods ALPHA is using the set of FORTRAN wrapped HBOOK routines provided by Anaphe LHC in hbook h Note that ALPHA already opens and closes this file 5 5 More about the ROOT interface The file defined in the HIST card is opened for ROOT by the driver By default ALPHA uses a tree called analysis At the end the tree and the file are closed Refer to the ROOT documentation to see how to place more objects in the file or how to directly edit the analysis tree 6 Event and Run informations Event and run information are available via dedicated classes To access the in formation first instantiate the classes from AlphaBanks AlphaBanks is passed by reference to the UserEvent routine The instance will be referred to as EventInfo AlEvent amp myevent EventInfo Event AlRun amp myrun EventInfo Run The use of references makes the code faster which is crucial in the UserEvent routine repeated Neven times The methods of these objects give access to the relevant information 6 1 AlEvent methods 11 6 1 AlEvent methods TimeInfo Time returns the event time Timelnfo is a utility clas
42. t a list of candidates Electron and muon idendification parameters are stored in two banks EIDT and MUID The first step is intended to get a set of candidates For electrons and muons this is a set of all the objects with a EIDT MUID entry The Lorentz vector is taken from the associated track For the taus this is a user generated set produced using the AtauJnQ method The list is thus obtained using e AlephCollection lt AlMuon gt muons bank muonPV e AlephCollection lt AlElec gt elecs bank elecPV 28 e AlephCollection lt AlEflw gt eflws bank eflwPV 5 vector lt float gt taidcuts theCardsReader gt TAIDcuts parameters Ycut scheme metric Evis AlephCollection lt AlTau gt taus eflws ATauJn taidcuts Evis AlephCollection lt AlTau gt tausP taus Pointers Note the tau specific step where the user produces a collection of pointers from the collection of candidates Standard parameters for AtauJn are Ycut 0 001 Cut used to build the jets Scheme 0 for EO ou 1 for E Metric 1 for Jade for Durham Evis Total energy If 0 is chosen the visible energy is used The standard usage is thus AtauJn ycut 1 1 with you 5 8 25 ELEP Filter the candidates The second step is the actual filtering using more precise cuts vector lt float gt muidcuts theCardsReader MUIDcuts vector lt float gt elidcuts theCardsReader gt ELIDcuts vector lt
43. t output files To override the defaults you can still redefine e ALPHACARDS the ALPHA 4 cards file e APPL OUT the driver session output file For example setenv ALPHACARDS mycards cards The software can be found on CVS The scripts define the CVSROOT variable so that you just have to checkout from the repository with co P r ALPHAPP 3 4 Applications The numbers indicate the version The last version can be found on the AL webpage This command will create an Applications directory contain ing all the code and some examples To compile an example just type make The executable file will be created in the Linux directory Execute the code by doing Linux acoplanar 10 This should produce a test root file To run the resulting code you have to provide the configuration cards file The default is ALPHA input This file is similar to the ALPHA one but with specific cards Available cards are described in the next section To change the compiled code just edit the GNUmakefile and replace the USER entry by what is needed If the code you want to build is coded in test3 cpp set USER test3 and run make again Three other test codes are available at startup e test1 cpp print the Lorentz vector of all Energy flows e test2 cpp select leptons and build jets Some screen outputs e test3 cpp look for the main vertex and compare it with the MC expectation Compute the b probability a
44. to use the ALPHA framework first some environment variables have to be set mydirectory is your work directory example afs cern ch aleph project database delaere should not mandatory contain any code setenv MYOODIR mydirectory setenv PRO afs cern ch aleph project database pro setenv DEV afs cern ch aleph project database dev setenv PRODB PRO db setenv DEVDB DEV db setenv ALPHACARDS alphat cards setenv ROOTSYS afs cern ch na49 library 4 ROOT new i386_redhat60 root setenv PATH ROOTSYS bin PATH setenv LD_LIBRARY_PATH ROOTSYS lib source DEV user_env csh alias roseinit source pttools Rose releases osf config csh alias roserun rose alias roseana analyzer alias insureinit source pttools Insure insure linux insure csh setenv CVSROOT pserver cerncvs pclhcb30 cern ch local alphat Some comments e Don t forget to set the MYOODIR field with your working directory e You may also have an already defined LD_LIBRARY_PATH Test it first with echo LD_LIBRARY_PATH and then use usually it has to be placed in the user s tcshrc file 2 Getting started 3 setenv LD_LIBRARY_PATH ROOTSYS lib LD LIBRARY PATH e if you already use ROOT you don t have to set the related variables The script sets some environment variables you need to access the database the Anaphe LHC tools ROOT utilities and the CVS repository The environnement script defines also standard ALPHA inpu
45. tus ReconstructorHeader ML calErrorFlags Readout BeamPosition BTriggerBitNames TriggerLevelControl TriggerOptions LuminosityParameters SicalLuminosityParameters HVfixes eTriggerLevelEventBitPatter TriggerLevelRegisterBitPattern PTriggerBitPattern MVDETbit NoHVbit ENELWinfo SicalTriggerX2 Triggerlevel1Threshold Triggerlevel1 Input bsicalTriggerSectorsAdc SicalTriggerConditions SicalRunConditions TriggerOldlevel 1 Threshold Triggerlevel1Adc 1 imulateSicalTrigger Apply Trigger JadeProtoJet BgetYTimesE2 operator findNextNeighbour BgetConstituants get4Momentum Scheme int DurhamProtoJet getYTimesE2 operator findNextNeighbour PgetConstituants Bget4Momentum Timelnfo msec int c int emin int aphour int pday int month int pyear int Timelnfo Timelnfo Poperator operator gt 41 Full Interactive protocol description 43 C Full Interactive protocol description This section will describe the present state of the protocol designed for an easy communi cation betwen ALPHA and any external software Up to now only AlVisu is using this feature This protocol is still allowed to change in the future in order to be more general or flexible The main features of t
46. ual 5 for more details As shown figure 1 the job is handled by the driver The user initialisation is called at the start of the job after the standard ALPHA initialisation There you can book histograms and or ntuples and perform other setup for the job The user event analysis is called for each event The returned value is used to select intersting events during an interactive session The user program termination is called before closing the output files and the BOS structure in memory 4 User routines Alpha Job initialisation input output initialisation database initialisation event initialisation alpha initialisation user initialisation Event processing database readout User Event Job termination event termination user termination alpha termination database termination input output termination Figure 1 The ALPHA job processing 8 5 Creating histograms and ntuples The standard histogram packages of ALPHA are HBOOK and ROOT The output file is given by the use of the HIST card in the alpha input cardfile using the syntax HIST file If the file given is a root file the ROOT package is initialized and used for output otherwise the file is given as input of the standard ALPHA routines so that an exch file is created following ALPHA rules A default system independent set of routines is provided to create a ntuple or a ROOT tree
47. ush_back 0 py output push_back 0 pz output push_back 0 ch SendMessage Code output Dummy test routine output clear output push_back 1 output push back 2 an eflow output push_back 10 E output push back 0 px output push back 10 py output push back 0 pz output push back 1 ch SendMessage Code output Dummy test routine output clear output push_back 0 output push_back 4 jet output push back 0 5 sin theta radius output push back 0 px output push back 0 py output push back 5 pz SendMessage Code output Dummy test routine virtual vector lt pair lt string float gt gt OptionsList f vector lt pair lt string float gt gt output pair lt string float gt mypair mypair first option 1 mypair second 10 REFERENCES 37 output push back mypair mypair first dummy 2 mypair second 3 1415 output push_back mypair return output void AlephExManager UserInit AlephRegisteredAction lt testclass gt mytestclass In this example nothing physical is done a set of messages is just build and send It will produce a track an eflow and a jet in AlVisu A more interesting example is the code written for the default actions event tracks leptons jets it can be found in the DefaultInteractiveActions cpp file References 1 http alephwww cern ch LIGHT alpha html 2 http www doxygen
48. ut sdtin of an external program This is the way AlVisu the ALEPH OO Visualisation Tool gets data from ALEPH databases 11 Interactive Mode 35 Internally there is an AlephInteractiveHandler which will send and receive messages in interactive mode This will not be discussed here There is a set of possible requests and the corresponding answers predefined in the driver in order to perform the basic data extraction of evente information tracks leptons and jets The full protocol is presented in annexe C In addition the user can define his own functions to call to produce a given result in interactive mode This is done by subclassing the AlephInteractiveHandler and by instanci ating the new class via the templated AlephRegistered Action T Shortly what you have to do e Create a new class deriving from AlephAbstractInteractiveFunction e Implement virtual string Name returning the name of the function e Implement virtual int Code returning a number larger than 500 e Implement virtual vector lt pair lt string float gt gt OptionsList returning a vector of pair option name default value e Implement virtual void Run vector lt float gt amp options AlphaBanks amp EventInfo the actual function UserEvent like e In UserInit instanciate AlephRegisteredAction lt T gt templated with your new class The Run method is called when needed with a vector of options and a reference to the AlphaBanks so that you c
49. vector from QVEC and the raw Lorentz vector from raw Gampeck info Usualy the corrected and the standard vectors are the same void UseCorrectedA4v void UseStandardA4V void UseRawA4V 0 Another set of routines is intended to return the Lorentz vector either from Gampeck or from QVEC HepLorentzVector GetCorrectedA4V const HepLorentzVector GetStandardA4V const HepLorentzVector GetRawA4V const Then various standard quality factors are provided Energy fraction in stack 1 or 2 float EnergyFractionInStack const int Energy fraction in the 4 central towers float EnergyFractionInCentralTowers Distance to the closest track cm float Isolation Storey flag int StoreyFlag Quality flag int QualityFlagO Finally the so called advanced quality factors are also interfaced Quality estimator 1 or 2 for photon float QualityEstimator const int 8 3 AlMCtruth the Monte Carlo truth information 17 moment 1 or 2 from CLMONS analysis float Moment const int PiO mass estmated from clmoms float Pi0Mass Expected fraction in 4 towers float ExpectedEnergyFractionInCentralTowers Geometrical correction float GeometricalCorrection Zero supression correction from Coradoc float ZeroSupression Probability to be a fake photon from Electromagnetic origin float FakeEcalProbability Probability to be a fake photon from Hadronic origin float FakeHcalProbab
50. vrtBase class 15 7 4 The QvrtBase class The QvrtBase class is very similar to the QvecBase The main methods e A4V returns the 4 vector By definition the energy of a vertex is 0 e ALEPHTYPE TYPE returns the object type float SYO 020 QEO QCTO QBETAO QGAMMA The meaning of those routines is the same as for AL PHA e int KVNQ returns the ALPHA vertex number e int KVNTYPE returns the vertex type e double QVCHIFO x of the vertex e HepSymMatrix QVEM the covariance matrix e float QVEM int i a covariance matrix element A vertex can of course be un locked The methods are the same as for QvecBase objects Lock unLock isLocked 8 Tracks and vectorial objects Here the additional methods specific to the different types of QvecBase derived objects are described 8 1 AlEflw the Enery Flow object The specific methods are EFLWTYPE GetEftype returns the Eflow type You can use the integer value as explained in the ALPHA manual or the corre sponding enumerated list EFLWTYPE Chargedtrack Electron Muon VOTrack Electromagnetic ECAL HCAL LCAL SICAL AlTrack getTrack returns the associated track 16 8 2 AlGamp the Gampec object A first set of routines is provided in order to select the Lorentz vector to use with the routines Possibilities are the corrected Lorentz vector from Gampeck the standard Lorentz
51. y are easy to apply The Aleph Collection is deriving from the STL vector In addition the looperase method makes the selection easier To get the collections just call the AlphaBanks methods AlephCollection lt AlGamp gt mygampeccollection EventInfo GampPV AlephCollection lt AlEflw gt myeflowcollection EventInfo EflwPV AlephCollection lt AlTrack gt mytrackcollection EventInfo TrackPV AlephCollection lt AlTrack gt mystdv0collection EventInfo stdVOPV AlephCollection lt AlTrack gt myvOtrackcollection EventInfo VOtrackPV AlephCollection lt AlTrack gt mylongv0collection EventInfo longVOPV AlephCollection lt AlMCtruth gt myMCcollection EventInfo MCtruthPV AlephCollection lt AlMuon gt mymuoncollection EventInfo MuonPV AlephCollection lt AlElec gt myeleccollection EventInfo ElecPV AlephCollection lt AlVertex gt myMCvertices EventInfo MCverticesPV AlephCollection lt AlVertex gt mysecvertices EventInfo SecVerticesPV AlVertex mymainvertex EventInfo MainVertexP Other methods returning the original vector of objects as a reference are avail able It should be noticed that in the latter case it is possible to modify the object a deleted track will definetly be lost Using the collections of pointers you can always retrieve a new full collection To loop on the collections there are two possibilities for AlephCollection lt AlTrack gt iterator itrack my
Download Pdf Manuals
Related Search