[u/mrichter/AliRoot.git] / EMCAL / doc / geometry.tex

\section{EMCAL geometry software}
(Marco Bregant, Alexandre Shabetai, Gustavo Conesa Balbastre)\\

This page is intended for a description of the EMCAL geometry and the methods to access it.  
{\it This is a very preliminary version that needs work.}

 \subsection {Classes description}

The EMCAL geometry is implemented in several classes : {\color{red} (right now very brief description, it should be completed) }
\begin{itemize}
\item AliEMCALGeometry: Steering geometry class.  No dependencies on STEER or EMCAL non geometry classes. Can be called during the analysis without loading all AliRoot classes.
\item AliEMCALEMCGeometry: Does the geometry initialization. Does all the definitions of the geometry (towers composition, size, Super Modules number ...)
\item AliEMCALGeoParams:  Class container of some of the geometry parameters so that it can be accessed everywhere in the EMCAL code, to avoid "magic numbers". Its use has to be propagated to all the code.
\item AliEMCALShishKebabTrd1Module: Here the modules are defined and the position of the modules in the local super module reference system is calculated
\end{itemize}

\subsection{Accessing the geometry}
One can get the geometry pointer in the following way:

%\begin{DDbox}{\linewidth}
\begin{lstlisting}
AliEMCALGeometry * geom =  AliEMCALGeometry::GetInstance(GeoName) ;
\end{lstlisting}
%\end{DDbox}

where "GeoName" is a string, the different options are specified in the next section. If you have a already generated simulation file and want to access the geometry used there, if the file {\it galice.root} is available:

\begin{DDbox}{\linewidth}
\begin{lstlisting}
AliRunLoader *rl = AliRunLoader::Open("galice.root",AliConfig::GetDefaultEventFolderName(),"read"); 
rl->LoadgAlice();//Needed to get geometry
AliEMCALLoader *emcalLoader = dynamic_cast<AliEMCALLoader*>(rl->GetDetectorLoader("EMCAL"));
AliRun * alirun = rl->GetAliRun();
AliEMCAL * emcal = (AliEMCAL*)alirun->GetDetector("EMCAL"); AliEMCALGeometry * geom = emcal->GetGeometry();
\end{lstlisting}
\end{DDbox}

In this case you might need the file {\it geometry.root} if you want to access to certain methods that require local to global position transformations. This file can be generated doing a simple simulation, it just contains the transformation matrix to go from global to local.

The way to load this file is:
%\begin{DDbox}{\linewidth}
\begin{lstlisting}
TGeoManager::Import("geometry.root");
\end{lstlisting}
%\end{DDbox}

The transformation matrices are also stored in the ESDs so if you do not load this file, you can have to load these matrices from the ESDs. They are also stored in the OADB.

 If you want to see different parameters used in the geometry printed (cells centers, distance to IP, etc), one just has to execute the method {\it PrintGeometry()}.
 
\subsection{Geometry configuration options}
Right now the following geometry options are implemented:
\begin{itemize}
\item EMCAL\_FIRSTYEARV1: 4 Super Modules, geometry  configuration for year 2010.
\item EMCAL\_COMPLETEV1: 10 Super Modules, geometry configuration for year 2011.
\item EMCAL\_COMPLETE12SMV1: 12 Super Modules (10+2$\times$1/3) full EMCal. Configuration for year 2012, although the 1/3 super-modules where not active.
\item EMCAL\_COMPLETE12SMV1\_DCAL: Full EMCal plus 6 DCal super-modules.
\item EMCAL\_COMPLETE12SMV1\_DCAL\_8SM: Full EMCal plus 6 DCal super-modules plus 2$\times$1/3 EMCal, geometry configuration for Run2 (years 2015-18).
\item EMCAL\_COMPLETE12SMV1\_DCAL\_DEV: Full EMCal plus  10 DCal super-modules (possible future configuration).
\end{itemize}

Other options exists but need to be removed as they {\bf should not be used} or that should be avoided:
\begin{itemize}
\item EMCAL\_PDC06: Old geometry, for reading old data (which do not exist anymore).
\item EMCAL\_WSU: Prototype geometry. 
\item EMCAL\_COMPLETE: 12 Super Modules (2 half Super Modules). Not correct geometry, use "V1" in name
\item EMCAL\_FIRSTYEAR: 4 Super Modules (year 2010). Not correct geometry, use "V1" in name.
\end{itemize}

By default, the geometry is loaded with the EMCAL\_COMPLETE12SMV1 configuration. For details on the implementation of the DCal geometry have a look here~\cite{DCalGeoOff}
 
 \subsection{Setting the geometry in simulations}

When configuring a simulation, a typical file {\it Config.C} is used. Inside this file, the detectors to be used in the generation are specified, and particularly EMCal is initialized in this way:

\begin{lstlisting}
AliEMCAL *EMCAL =  = new AliEMCALv2("EMCAL", TString GeoName, Bool_t checkGeoRun);
\end{lstlisting}

Where:
\begin{itemize}
  \item {\it AliEMCAL} is the main steering class for the simulation. The derived classes {\t AliEMCALvX}:
    \begin{itemize}
    \item {\it AliEMCALv0} does the geometry initialization, materials creation etc.
    \item {\it AliEMCALv1} derives from {\it v0}, DO NOT USE for simulation, originally it was meant for fast simulations, it does not generate hits.
    \item {\it AliEMCALv2} derives from {\it v1}, USE for simulation. It does all the particle propagation in the EMCal material
\end{itemize}
  \item {\it TString GeoName}: Geometry names listed in the previous section
  \item {\it Bool\_t checkGeoRun}: Bool that activates the geometry initialization depending on the run number. Since EMCal geometry changed over time, to avoid mistakes in the configuration files of simulations anchored to particular run numbers,  by default this is set to TRUE and the name set in the initialization of AliEMCAL is not taken into account but the one corresponding to this run.
\end{itemize}

 
 \subsection{Mapping}
The tower row/column mapping online and offline follows the alice numbering convention. Figures~\ref{fig:Map1} to \ref{fig:Map2} display the position of the super modules from different points of view and the position of the tower index in them.

\begin{figure}[ht]
\begin{center}
\includegraphics[width=1.0\textwidth]{figures/EMCALMap0_0.pdf}
\end{center}
\caption{\label{fig:Map1}Position of the super modules}
\end{figure}

\begin{figure}[ht]
\begin{center}
\includegraphics[width=1.0\textwidth]{figures/EMCALMap1.pdf}
\end{center}
\caption{\label{fig:Map0}EMCal seen from the magnet side with 4 SMs.}
\end{figure}


\begin{figure}[ht]
\begin{center}
\includegraphics[width=1.0\textwidth]{figures/EMCALMap2.pdf}
\end{center}
\caption{\label{fig:Map2}EMCal geometrical numbering.}
\end{figure}


\subsection{Tower index transformation methods}
\subsubsection{Absolute tower ID to Row/Column index}

Each EMCAL supermodule is composed of 24x48 towers (phi,eta), grouped in 4x4 modules. Each tower (even each module) has a unique number assigned, called in the code "absolute ID" number (absId). This number can be transformed into a row (phi direction) or column (eta direction) index. The procedure to go from the absId to the (row, col) formulation or viceversa is as follow:

\begin{itemize}
\item From absId to col-row: 

\begin{DDbox}{\linewidth}
\begin{lstlisting}
Int_t nSupMod, nModule, nIphi, nIeta, iphi, ieta;
//Check if this absId exists
if(!CheckAbsCellId(absId)) return kFALSE;
// Get from the absId the super module number, the module number and the eta-phi index (0 or 1) in the module
GetCellIndex(absId, nSupMod, nModule, nIphi, nIeta);
// Get from the  the super module number, the module number and the eta-phi index (0 or 1) in the module the tower row (iphi) and column (ieta) 
GetCellPhiEtaIndexInSModule(nSupMod,nModule,nIphi,nIeta, iphi, ieta); 
\end{lstlisting}
\end{DDbox}

\item From col-row to absId, following the same notation as above: 

\begin{DDbox}{\linewidth}
\begin{lstlisting}

absid = GetAbsCellIdFromCellIndexes(nSupMode, iphi, ieta);
\end{lstlisting}
\end{DDbox}

or

\begin{DDbox}{\linewidth}
\begin{lstlisting}

absid = GetAbsCellId(nSupMod, nModule, nIphi, nIeta);
\end{lstlisting}
\end{DDbox}

\item Other interesting method is

\begin{DDbox}{\linewidth}
\begin{lstlisting}

Int_t   GetSuperModuleNumber(Int_t absId)  
\end{lstlisting}
\end{DDbox}
\end{itemize}

\subsection{Tower index to local / global reference system position}
\subsubsection{Local coordinates}

To correlate the tower index and its position in local coordinates, the following methods are available:

\begin{DDbox}{\linewidth}
\begin{lstlisting}
  Bool_t  AliEMCALGeoUtils::RelPosCellInSModule(Int_t absId, Double_t &xr, Double_t &yr, Double_t &zr) const;

  Bool_t  AliEMCALGeoUtils::RelPosCellInSModule(Int_t absId, Double_t loc[3]) const;

  Bool_t  AliEMCALGeoUtils::RelPosCellInSModule(Int_t absId, TVector3 &vloc) const;
\end{lstlisting}
\end{DDbox}

To which the input is the absId and the output are the coordinates of the center of towers in the local coordinates of the Super Module. This method gets the column and row index of the cell from the absId, independently of the Super Module (like above), and gets the center of the cell from 3 arrays (x,y,z) filled with such quantities. Such central positions are calculated during the initialization of the geometry, where the arrys are filled, in the method :

\begin{DDbox}{\linewidth}
\begin{lstlisting}

  AliEMCALGeoUtils::CreateListOfTrd1Modules()
\end{lstlisting}
\end{DDbox}

{\color{red}<<<<<Someone else should explain how it works>>>>>}

 
In case we calculate the cluster position, things are a bit different.

 
{\color{red}<<<<< This explanation should go to the clusterization section>>>>}

 
This is done in 

\begin{DDbox}{\linewidth}
\begin{lstlisting}
void AliEMCALRecPoint::EvalLocalPosition()
\end{lstlisting}
\end{DDbox}

First we calculate the cell position with the  method 

\begin{DDbox}{\linewidth}
\begin{lstlisting}
AliEMCALGeometry::RelPosCellInSModule(Int_t absId, Int_t maxAbsId, Double_t tmax, Double_t &xr, Double_t &yr, Double_t &zr)
\end{lstlisting}
\end{DDbox}

The calculation of the cell position done here is different in the "x-z" but the same in "y". 

 
{\color{red}<<<<< <<<<<Someone else should explain how it works>>>>>}

 
In this particular case the position calculation per tower depends on the position of the maximum cell, and the sum of the energy of the cells of the cluster. The maximum depth (tmax) is calculated with the method

\begin{DDbox}{\linewidth}
\begin{lstlisting}
  Double\t AliEMCALRecPoint::TmaxInCm(const Double_t e){

      //e: energy sum of cells

static Double_t ca = 4.82;// shower max parameter - first guess; ca=TMath::Log(1000./8.07)

      static Double_t x0 = 1.23; // radiation lenght (cm)

      static Double_t tmax = 0.; // position of electromagnetic shower max in cm

      tmax = TMath::Log(e) + ca+0.5;

      tmax *= x0; // convert to cm

  }
\end{lstlisting}
\end{DDbox}

After the cells position of the cluster is accessed, the position of the cluster is calculated averaging the cell positions with a logarithmic weight:

\begin{DDbox}{\linewidth}
\begin{lstlisting}
  w(cell i) = TMath::Max( 0., logWeight + TMath::Log( energy[cell i] / summed_cluster_cell_energy ));
\end{lstlisting}
\end{DDbox}

where the logWeight was chosen to be 4.5 (this value was taken from PHOS, never optimized as far as I know)

So in the end the position,  is 

\begin{DDbox}{\linewidth}
\begin{lstlisting}
f = Sum(f(i) * w(i))/Sum(w(i))
\end{lstlisting}
\end{DDbox}

where f=x,y,z.

 
 \subsubsection{Global coordinates}
 

To transform from local to global we have the methods

\begin{DDbox}{\linewidth}
\begin{lstlisting}

  void GetGlobal(const Double_t *loc, Double_t *glob, int ind) const;

  void GetGlobal(const TVector3 &vloc, TVector3 &vglob, int ind) const;

  void GetGlobal(Int_t absId, Double_t glob[3]) const;

  void GetGlobal(Int_t absId, TVector3 &vglob) const;
\end{lstlisting}
\end{DDbox}

 
 These methods take the local coordinates and transform them into global coordinates using the transformation matrix of the Super Module.

\begin{DDbox}{\linewidth}
\begin{lstlisting}
 

TGeoHMatrix* m = GetMatrixForSuperModule(nSupMod);

if(m) m->LocalToMaster(loc, glob);

\end{lstlisting}
\end{DDbox}
 

 GetGlobal is called in the following useful methods in the geometry class:

 
\begin{itemize}
\item Return the eta and phi angular position of the cell from the AbsId

\begin{DDbox}{\linewidth}
\begin{lstlisting}
void    EtaPhiFromIndex(Int_t absId, Double_t &eta, Double_t &phi) const;
void    EtaPhiFromIndex(Int_t absId, Float_t &eta, Float_t &phi) const;
\end{lstlisting}
\end{DDbox}
\item Print information of the cells. For "pri>0" returns more information. "tit" has not much use, this value is printed.

\begin{DDbox}{\linewidth}
\begin{lstlisting}
void PrintCellIndexes(Int_t absId, int pri, const char *tit)
\end{lstlisting}
\end{DDbox}
\end{itemize} 

\subsection{Geometry Alignment}
AliRoot contains a frame for the correction of the misplacement of geometry objects with respect to the ideal positions which are kept in the STEER/ directory of the following classes:

\begin{DDbox}{\linewidth}
\begin{lstlisting}
AliAlignObj  
AliAlignObjMatrix  
AliAlignObjParams  
AliAlignmentTracks
\end{lstlisting}
\end{DDbox}

The class AliEMCALSurvey creates the corrections to the alignable objects.  The class AliEMCALSurvey was established to take the survey parameters from OCDB, calculate the shift in position of the center of the end faces of the supermodules from the nominal position, and convert this to a transformation matrix for each supermodule which is applied to correct the global position of the supermodules.  All calculations of global positions would then use these corrected supermodule positions to determine their locations within the ALICE global coordinate system.
Commit	Line	Data
552957f3	1	\section{EMCAL geometry software}
552957f3	2	(Marco Bregant, Alexandre Shabetai, Gustavo Conesa Balbastre)\\
02582a78	3
f5c540cb	4	This page is intended for a description of the EMCAL geometry and the methods to access it.
f5c540cb	5	{\it This is a very preliminary version that needs work.}
02582a78	6
f5c540cb	7	\subsection {Classes description}
	8
	9	The EMCAL geometry is implemented in several classes : {\color{red} (right now very brief description, it should be completed) }
02582a78	10	\begin{itemize}
552957f3	11	\item AliEMCALGeometry: Steering geometry class. No dependencies on STEER or EMCAL non geometry classes. Can be called during the analysis without loading all AliRoot classes.
02582a78	12	\item AliEMCALEMCGeometry: Does the geometry initialization. Does all the definitions of the geometry (towers composition, size, Super Modules number ...)
	13	\item AliEMCALGeoParams: Class container of some of the geometry parameters so that it can be accessed everywhere in the EMCAL code, to avoid "magic numbers". Its use has to be propagated to all the code.
	14	\item AliEMCALShishKebabTrd1Module: Here the modules are defined and the position of the modules in the local super module reference system is calculated
	15	\end{itemize}
	16
f5c540cb	17	\subsection{Accessing the geometry}
552957f3	18	One can get the geometry pointer in the following way:
	19
	20	%\begin{DDbox}{\linewidth}
	21	\begin{lstlisting}
	22	AliEMCALGeometry * geom = AliEMCALGeometry::GetInstance(GeoName) ;
	23	\end{lstlisting}
	24	%\end{DDbox}
	25
	26	where "GeoName" is a string, the different options are specified in the next section. If you have a already generated simulation file and want to access the geometry used there, if the file {\it galice.root} is available:
f5c540cb	27
	28	\begin{DDbox}{\linewidth}
	29	\begin{lstlisting}
	30	AliRunLoader *rl = AliRunLoader::Open("galice.root",AliConfig::GetDefaultEventFolderName(),"read");
	31	rl->LoadgAlice();//Needed to get geometry
552957f3	32	AliEMCALLoader emcalLoader = dynamic_cast<AliEMCALLoader>(rl->GetDetectorLoader("EMCAL"));
f5c540cb	33	AliRun * alirun = rl->GetAliRun();
f5c540cb	34	AliEMCAL * emcal = (AliEMCAL)alirun->GetDetector("EMCAL"); AliEMCALGeometry geom = emcal->GetGeometry();
f5c540cb	35	\end{lstlisting}
f5c540cb	36	\end{DDbox}
552957f3	37
552957f3	38	In this case you might need the file {\it geometry.root} if you want to access to certain methods that require local to global position transformations. This file can be generated doing a simple simulation, it just contains the transformation matrix to go from global to local.
02582a78	39
f5c540cb	40	The way to load this file is:
	41	%\begin{DDbox}{\linewidth}
	42	\begin{lstlisting}
	43	TGeoManager::Import("geometry.root");
	44	\end{lstlisting}
	45	%\end{DDbox}
02582a78	46
552957f3	47	The transformation matrices are also stored in the ESDs so if you do not load this file, you can have to load these matrices from the ESDs. They are also stored in the OADB.
02582a78	48
552957f3	49	If you want to see different parameters used in the geometry printed (cells centers, distance to IP, etc), one just has to execute the method {\it PrintGeometry()}.
02582a78	50
02582a78	51	\subsection{Geometry configuration options}
f5c540cb	52	Right now the following geometry options are implemented:
02582a78	53	\begin{itemize}
eaccbd9e	54	\item EMCAL\_FIRSTYEARV1: 4 Super Modules, geometry configuration for year 2010.
	55	\item EMCAL\_COMPLETEV1: 10 Super Modules, geometry configuration for year 2011.
	56	\item EMCAL\_COMPLETE12SMV1: 12 Super Modules (10+2$\times$1/3) full EMCal. Configuration for year 2012, although the 1/3 super-modules where not active.
486ebd44	57	\item EMCAL\_COMPLETE12SMV1\_DCAL: Full EMCal plus 6 DCal super-modules.
eaccbd9e	58	\item EMCAL\_COMPLETE12SMV1\_DCAL\_8SM: Full EMCal plus 6 DCal super-modules plus 2$\times$1/3 EMCal, geometry configuration for Run2 (years 2015-18).
486ebd44	59	\item EMCAL\_COMPLETE12SMV1\_DCAL\_DEV: Full EMCal plus 10 DCal super-modules (possible future configuration).
02582a78	60	\end{itemize}
02582a78	61
552957f3	62	Other options exists but need to be removed as they {\bf should not be used} or that should be avoided:
02582a78	63	\begin{itemize}
f5c540cb	64	\item EMCAL\_PDC06: Old geometry, for reading old data (which do not exist anymore).
02582a78	65	\item EMCAL\_WSU: Prototype geometry.
eaccbd9e	66	\item EMCAL\_COMPLETE: 12 Super Modules (2 half Super Modules). Not correct geometry, use "V1" in name
eaccbd9e	67	\item EMCAL\_FIRSTYEAR: 4 Super Modules (year 2010). Not correct geometry, use "V1" in name.
02582a78	68	\end{itemize}
02582a78	69
486ebd44	70	By default, the geometry is loaded with the EMCAL\_COMPLETE12SMV1 configuration. For details on the implementation of the DCal geometry have a look here~\cite{DCalGeoOff}
02582a78	71
552957f3	72	\subsection{Setting the geometry in simulations}
	73
	74	When configuring a simulation, a typical file {\it Config.C} is used. Inside this file, the detectors to be used in the generation are specified, and particularly EMCal is initialized in this way:
	75
	76	\begin{lstlisting}
	77	AliEMCAL *EMCAL = = new AliEMCALv2("EMCAL", TString GeoName, Bool_t checkGeoRun);
	78	\end{lstlisting}
	79
	80	Where:
	81	\begin{itemize}
	82	\item {\it AliEMCAL} is the main steering class for the simulation. The derived classes {\t AliEMCALvX}:
	83	\begin{itemize}
	84	\item {\it AliEMCALv0} does the geometry initialization, materials creation etc.
	85	\item {\it AliEMCALv1} derives from {\it v0}, DO NOT USE for simulation, originally it was meant for fast simulations, it does not generate hits.
	86	\item {\it AliEMCALv2} derives from {\it v1}, USE for simulation. It does all the particle propagation in the EMCal material
	87	\end{itemize}
	88	\item {\it TString GeoName}: Geometry names listed in the previous section
	89	\item {\it Bool\_t checkGeoRun}: Bool that activates the geometry initialization depending on the run number. Since EMCal geometry changed over time, to avoid mistakes in the configuration files of simulations anchored to particular run numbers, by default this is set to TRUE and the name set in the initialization of AliEMCAL is not taken into account but the one corresponding to this run.
	90	\end{itemize}
	91
	92
02582a78	93	\subsection{Mapping}
f5c540cb	94	The tower row/column mapping online and offline follows the alice numbering convention. Figures~\ref{fig:Map1} to \ref{fig:Map2} display the position of the super modules from different points of view and the position of the tower index in them.
02582a78	95
	96	\begin{figure}[ht]
	97	\begin{center}
3ae81835	98	\includegraphics[width=1.0\textwidth]{figures/EMCALMap0_0.pdf}
02582a78	99	\end{center}
6ebd8b1a	100	\caption{\label{fig:Map1}Position of the super modules}
02582a78	101	\end{figure}
	102
	103	\begin{figure}[ht]
	104	\begin{center}
3ae81835	105	\includegraphics[width=1.0\textwidth]{figures/EMCALMap1.pdf}
02582a78	106	\end{center}
6ebd8b1a	107	\caption{\label{fig:Map0}EMCal seen from the magnet side with 4 SMs.}
02582a78	108	\end{figure}
	109
	110
	111	\begin{figure}[ht]
	112	\begin{center}
3ae81835	113	\includegraphics[width=1.0\textwidth]{figures/EMCALMap2.pdf}
02582a78	114	\end{center}
6ebd8b1a	115	\caption{\label{fig:Map2}EMCal geometrical numbering.}
02582a78	116	\end{figure}
	117
	118
	119	\subsection{Tower index transformation methods}
	120	\subsubsection{Absolute tower ID to Row/Column index}
02582a78	121
f5c540cb	122	Each EMCAL supermodule is composed of 24x48 towers (phi,eta), grouped in 4x4 modules. Each tower (even each module) has a unique number assigned, called in the code "absolute ID" number (absId). This number can be transformed into a row (phi direction) or column (eta direction) index. The procedure to go from the absId to the (row, col) formulation or viceversa is as follow:
02582a78	123
f5c540cb	124	\begin{itemize}
f5c540cb	125	\item From absId to col-row:
02582a78	126
f5c540cb	127	\begin{DDbox}{\linewidth}
f5c540cb	128	\begin{lstlisting}
552957f3	129	Int_t nSupMod, nModule, nIphi, nIeta, iphi, ieta;
f5c540cb	130	//Check if this absId exists
02582a78	131	if(!CheckAbsCellId(absId)) return kFALSE;
02582a78	132	// Get from the absId the super module number, the module number and the eta-phi index (0 or 1) in the module
02582a78	133	GetCellIndex(absId, nSupMod, nModule, nIphi, nIeta);
02582a78	134	// Get from the the super module number, the module number and the eta-phi index (0 or 1) in the module the tower row (iphi) and column (ieta)
02582a78	135	GetCellPhiEtaIndexInSModule(nSupMod,nModule,nIphi,nIeta, iphi, ieta);
f5c540cb	136	\end{lstlisting}
f5c540cb	137	\end{DDbox}
02582a78	138
f5c540cb	139	\item From col-row to absId, following the same notation as above:
	140
	141	\begin{DDbox}{\linewidth}
	142	\begin{lstlisting}
02582a78	143
02582a78	144	absid = GetAbsCellIdFromCellIndexes(nSupMode, iphi, ieta);
f5c540cb	145	\end{lstlisting}
f5c540cb	146	\end{DDbox}
02582a78	147
	148	or
	149
f5c540cb	150	\begin{DDbox}{\linewidth}
	151	\begin{lstlisting}
	152
02582a78	153	absid = GetAbsCellId(nSupMod, nModule, nIphi, nIeta);
f5c540cb	154	\end{lstlisting}
f5c540cb	155	\end{DDbox}
02582a78	156
f5c540cb	157	\item Other interesting method is
02582a78	158
f5c540cb	159	\begin{DDbox}{\linewidth}
f5c540cb	160	\begin{lstlisting}
02582a78	161
552957f3	162	Int_t GetSuperModuleNumber(Int_t absId)
f5c540cb	163	\end{lstlisting}
	164	\end{DDbox}
	165	\end{itemize}
02582a78	166
	167	\subsection{Tower index to local / global reference system position}
	168	\subsubsection{Local coordinates}
02582a78	169
f5c540cb	170	To correlate the tower index and its position in local coordinates, the following methods are available:
	171
	172	\begin{DDbox}{\linewidth}
	173	\begin{lstlisting}
552957f3	174	Bool_t AliEMCALGeoUtils::RelPosCellInSModule(Int_t absId, Double_t &xr, Double_t &yr, Double_t &zr) const;
02582a78	175
552957f3	176	Bool_t AliEMCALGeoUtils::RelPosCellInSModule(Int_t absId, Double_t loc[3]) const;
02582a78	177
552957f3	178	Bool_t AliEMCALGeoUtils::RelPosCellInSModule(Int_t absId, TVector3 &vloc) const;
f5c540cb	179	\end{lstlisting}
	180	\end{DDbox}
	181
	182	To which the input is the absId and the output are the coordinates of the center of towers in the local coordinates of the Super Module. This method gets the column and row index of the cell from the absId, independently of the Super Module (like above), and gets the center of the cell from 3 arrays (x,y,z) filled with such quantities. Such central positions are calculated during the initialization of the geometry, where the arrys are filled, in the method :
02582a78	183
f5c540cb	184	\begin{DDbox}{\linewidth}
f5c540cb	185	\begin{lstlisting}
02582a78	186
02582a78	187	AliEMCALGeoUtils::CreateListOfTrd1Modules()
f5c540cb	188	\end{lstlisting}
f5c540cb	189	\end{DDbox}
02582a78	190
f5c540cb	191	{\color{red}<<<<<Someone else should explain how it works>>>>>}
02582a78	192
	193
	194
	195	In case we calculate the cluster position, things are a bit different.
	196
	197
	198
f5c540cb	199	{\color{red}<<<<< This explanation should go to the clusterization section>>>>}
02582a78	200
	201
	202
	203	This is done in
	204
f5c540cb	205	\begin{DDbox}{\linewidth}
f5c540cb	206	\begin{lstlisting}
02582a78	207	void AliEMCALRecPoint::EvalLocalPosition()
f5c540cb	208	\end{lstlisting}
f5c540cb	209	\end{DDbox}
02582a78	210
	211	First we calculate the cell position with the method
	212
f5c540cb	213	\begin{DDbox}{\linewidth}
f5c540cb	214	\begin{lstlisting}
552957f3	215	AliEMCALGeometry::RelPosCellInSModule(Int_t absId, Int_t maxAbsId, Double_t tmax, Double_t &xr, Double_t &yr, Double_t &zr)
f5c540cb	216	\end{lstlisting}
f5c540cb	217	\end{DDbox}
02582a78	218
	219	The calculation of the cell position done here is different in the "x-z" but the same in "y".
	220
	221
	222
f5c540cb	223	{\color{red}<<<<< <<<<<Someone else should explain how it works>>>>>}
02582a78	224
	225
	226
	227
	228	In this particular case the position calculation per tower depends on the position of the maximum cell, and the sum of the energy of the cells of the cluster. The maximum depth (tmax) is calculated with the method
	229
f5c540cb	230	\begin{DDbox}{\linewidth}
f5c540cb	231	\begin{lstlisting}
552957f3	232	Double\t AliEMCALRecPoint::TmaxInCm(const Double_t e){
02582a78	233
	234	//e: energy sum of cells
	235
552957f3	236	static Double_t ca = 4.82;// shower max parameter - first guess; ca=TMath::Log(1000./8.07)
02582a78	237
552957f3	238	static Double_t x0 = 1.23; // radiation lenght (cm)
02582a78	239
552957f3	240	static Double_t tmax = 0.; // position of electromagnetic shower max in cm
02582a78	241
	242	tmax = TMath::Log(e) + ca+0.5;
	243
	244	tmax *= x0; // convert to cm
	245
	246	}
f5c540cb	247	\end{lstlisting}
f5c540cb	248	\end{DDbox}
02582a78	249
f5c540cb	250	After the cells position of the cluster is accessed, the position of the cluster is calculated averaging the cell positions with a logarithmic weight:
02582a78	251
f5c540cb	252	\begin{DDbox}{\linewidth}
f5c540cb	253	\begin{lstlisting}
552957f3	254	w(cell i) = TMath::Max( 0., logWeight + TMath::Log( energy[cell i] / summed_cluster_cell_energy ));
f5c540cb	255	\end{lstlisting}
f5c540cb	256	\end{DDbox}
02582a78	257
	258	where the logWeight was chosen to be 4.5 (this value was taken from PHOS, never optimized as far as I know)
	259
	260	So in the end the position, is
	261
f5c540cb	262	\begin{DDbox}{\linewidth}
f5c540cb	263	\begin{lstlisting}
02582a78	264	f = Sum(f(i) * w(i))/Sum(w(i))
f5c540cb	265	\end{lstlisting}
f5c540cb	266	\end{DDbox}
02582a78	267
	268	where f=x,y,z.
	269
	270
	271
	272	\subsubsection{Global coordinates}
	273
	274
	275	To transform from local to global we have the methods
	276
f5c540cb	277	\begin{DDbox}{\linewidth}
	278	\begin{lstlisting}
	279
552957f3	280	void GetGlobal(const Double_t loc, Double_t glob, int ind) const;
02582a78	281
552957f3	282	void GetGlobal(const TVector3 &vloc, TVector3 &vglob, int ind) const;
02582a78	283
552957f3	284	void GetGlobal(Int_t absId, Double_t glob[3]) const;
02582a78	285
552957f3	286	void GetGlobal(Int_t absId, TVector3 &vglob) const;
f5c540cb	287	\end{lstlisting}
f5c540cb	288	\end{DDbox}
02582a78	289
	290
	291
	292	These methods take the local coordinates and transform them into global coordinates using the transformation matrix of the Super Module.
	293
f5c540cb	294	\begin{DDbox}{\linewidth}
f5c540cb	295	\begin{lstlisting}
02582a78	296
	297
	298	TGeoHMatrix* m = GetMatrixForSuperModule(nSupMod);
	299
	300	if(m) m->LocalToMaster(loc, glob);
	301
f5c540cb	302	\end{lstlisting}
f5c540cb	303	\end{DDbox}
02582a78	304
	305
	306	GetGlobal is called in the following useful methods in the geometry class:
	307
	308
f5c540cb	309	\begin{itemize}
f5c540cb	310	\item Return the eta and phi angular position of the cell from the AbsId
02582a78	311
f5c540cb	312	\begin{DDbox}{\linewidth}
f5c540cb	313	\begin{lstlisting}
552957f3	314	void EtaPhiFromIndex(Int_t absId, Double_t &eta, Double_t &phi) const;
552957f3	315	void EtaPhiFromIndex(Int_t absId, Float_t &eta, Float_t &phi) const;
f5c540cb	316	\end{lstlisting}
	317	\end{DDbox}
	318	\item Print information of the cells. For "pri>0" returns more information. "tit" has not much use, this value is printed.
	319
	320	\begin{DDbox}{\linewidth}
	321	\begin{lstlisting}
552957f3	322	void PrintCellIndexes(Int_t absId, int pri, const char *tit)
f5c540cb	323	\end{lstlisting}
	324	\end{DDbox}
	325	\end{itemize}
02582a78	326
02582a78	327	\subsection{Geometry Alignment}
f5c540cb	328	AliRoot contains a frame for the correction of the misplacement of geometry objects with respect to the ideal positions which are kept in the STEER/ directory of the following classes:
	329
	330	\begin{DDbox}{\linewidth}
	331	\begin{lstlisting}
	332	AliAlignObj
	333	AliAlignObjMatrix
	334	AliAlignObjParams
	335	AliAlignmentTracks
	336	\end{lstlisting}
	337	\end{DDbox}
	338
	339	The class AliEMCALSurvey creates the corrections to the alignable objects. The class AliEMCALSurvey was established to take the survey parameters from OCDB, calculate the shift in position of the center of the end faces of the supermodules from the nominal position, and convert this to a transformation matrix for each supermodule which is applied to correct the global position of the supermodules. All calculations of global positions would then use these corrected supermodule positions to determine their locations within the ALICE global coordinate system.