class AliTPCcalibDButil: public TObject

 Default ctor

 dtor

 Update pointers from calibDB

 Process the CE data for this run
 the return TVectorD arrays contian the results of the fit
 noutliersCE contains the number of pads marked as outliers,
   not including masked and edge pads

 Calculate statistical information from the CE graphs for drift time and charge

 process noise data
 vNoiseMean/RMS contains the Mean/RMS noise of the complete TPC [0], IROCs only [1],
    OROCs small pads [2] and OROCs large pads [3]
 vNoiseMean/RMSsenRegions constains the same information, but only for the sensitive regions (edge pads, corners, IROC spot)
 nonMaskedZero contains the number of pads which show zero noise and were not masked. This might indicate an error

 process QA data

 vQaOcc/Qtot/Qmax contains the Mean occupancy/Qtot/Qmax for each sector

 Process the Pulser information
 vMeanTime:     pulser mean time position in IROC-A, IROC-C, OROC-A, OROC-C

 Get Values from ALTRO configuration data

 Proces Goofie values, return statistical information of the currently set goofieArray
 The meaning of the entries are given below

  1       TPC_ANODE_I_A00_STAT
  2       TPC_DVM_CO2
  3       TPC_DVM_DriftVelocity
  4       TPC_DVM_FCageHV
  5       TPC_DVM_GainFar
  6       TPC_DVM_GainNear
  7       TPC_DVM_N2
  8       TPC_DVM_NumberOfSparks
  9       TPC_DVM_PeakAreaFar
  10      TPC_DVM_PeakAreaNear
  11      TPC_DVM_PeakPosFar
  12      TPC_DVM_PeakPosNear
  13      TPC_DVM_PickupHV
  14      TPC_DVM_Pressure
  15      TPC_DVM_T1_Over_P
  16      TPC_DVM_T2_Over_P
  17      TPC_DVM_T_Over_P
  18      TPC_DVM_TemperatureS1

 check the variations of the pedestal data to the reference pedestal data
 thresholds are 0.5, 1.0, 1.5 and 2 timebins respectively.

 check the variations of the noise data to the reference noise data
 thresholds are 5, 10, 15 and 20 percent respectively.

 check the variations of the pulserQmean data to the reference pulserQmean data: pulserQdeviations
 thresholds are .5, 1, 5 and 10 percent respectively.

 Update the outlier map of the pulser data

 Update the outlier map of the pulser reference data

 Create a map that contains outliers from the Pulser calibration data.
 The outliers include masked channels, edge pads and pads with
   too large timing and charge variations.
 fNpulserOutliers is the number of outliers in the Pulser calibration data.
   those do not contain masked and edge pads

 Create pad time0 object from pulser and/or CE data, depending on the selected model
 Model 0: normalise each readout chamber to its mean, outlier cutted, only Pulser
 Model 1: normalise IROCs/OROCs of each readout side to its mean, only Pulser
 Model 2: use CE data and a combination CE fit + pulser in the outlier regions.

 In case model 2 is invoked - gy arival time gradient is also returned

 GetMeanAlto information

 load cal pad objects form the reference file

 set reference data from OCDB Reference map

 TObjArray object type case
 find 'objName' in 'arr' cast is to a calPad and store it in 'pad'

 AliTPCCalPad object type case
 cast object to a calPad and store it in 'pad'

 set altro masked channel map for 'cdbPath'

 Get Reference map

 check whether a reference cdb entry has changed

 get the reference AliCDBEntry for 'cdbPath'

 Get reference run number for the specified OCDB path

 Get reference run number for the specified OCDB path

 Author:  marian.ivanov@cern.ch

 Create outlier map for CE study
 Parameters:
  Return value - outlyer map
  noutlyersCE  - number of outlyers
  minSignal    - minimal total Q signal
  cutRMSMin    - minimal width of the signal in respect to the median
  cutRMSMax    - maximal width of the signal in respect to the median
  cutMaxDistT  - maximal deviation from time median per chamber

 Outlyers criteria:
 0. Exclude masked pads
 1. Exclude first two rows in IROC and last two rows in OROC
 2. Exclude edge pads
 3. Exclude channels with too large variations
 4. Exclude pads with too small signal
 5. Exclude signal with outlyers RMS
 6. Exclude channels to far from the chamber median

 Author: marian.ivanov@cern.ch

 Create outlier map for Pulser
 Parameters:
  Return value     - outlyer map
  noutlyersPulser  - number of outlyers
  cutTime          - absolute cut - distance to the median of chamber
  cutnRMSQ         - nsigma cut from median  q distribution per chamber
  cutnRMSrms       - nsigma cut from median  rms distribution
 Outlyers criteria:
 0. Exclude masked pads
 1. Exclude time outlyers (default 3 time bins)
 2. Exclude q outlyers    (default 5 sigma)
 3. Exclude rms outlyers  (default 5 sigma)

 Author : Marian Ivanov
 Create pad time0 correction map using information from the CE and from pulser


 Return PadTime0 to be used for time0 relative alignment
 if dump file specified intermediat results are dumped to the fiel and can be visualized
 using $ALICE_ROOT/TPC/script/gui application

 fitResultsA - fitParameters A side
 fitResultsC - fitParameters C side
 chi2A       - chi2/ndf for A side (assuming error 1 time bin)
 chi2C       - chi2/ndf for C side (assuming error 1 time bin)


 Algorithm:
 1. Find outlier map for CE
 2. Find outlier map for Pulser
 3. Replace outlier by median at given sector  (median without outliers)
 4. Substract from the CE data pulser
 5. Fit the CE with formula
    5.1) (IROC-OROC) offset
    5.2) gx
    5.3) gy
    5.4) (lx-xmid)
    5.5) (IROC-OROC)*(lx-xmid)
    5.6) (ly/lx)^2
 6. Substract gy fit dependence from the CE data
 7. Add pulser back to CE data
 8. Replace outliers by fit value - median of diff per given chamber -GY fit
 9. return CE data

 Time0 <= padCE = padCEin  -padCEfitGy  - if not outlier
 Time0 <= padCE = padFitAll-padCEfitGy  - if outlier

 find the closest point to xref  in x  direction
 return dx and value

 Get the correction of the trigger offset
 combining information from the laser track calibration
 and from cosmic calibration

 run       - run number
 timeStamp - tim stamp in seconds
 deltaT    - integration period to calculate offset
 deltaTLaser -max validity of laser data
 valType   - 0 - median, 1- mean

 Integration vaues are just recomendation - if not possible to get points
 automatically increase the validity by factor 2
 (recursive algorithm until one month of data taking)

 Get the correction of the drift velocity
 combining information from the laser track calibration
 and from cosmic calibration

 dist      - return value - distance to closest point in graph
 run       - run number
 timeStamp - tim stamp in seconds
 deltaT    - integration period to calculate time0 offset
 deltaTLaser -max validity of laser data
 valType   - 0 - median, 1- mean

 Integration vaues are just recomendation - if not possible to get points
 automatically increase the validity by factor 2
 (recursive algorithm until one month of data taking)

 Create a default name for the gui file

 Create a gui reference tree
 if dirname and filename are empty default values will be used
 this is the recommended way of using this function
 it allows to check whether a file with the given run validity alredy exists

 Get the correction of the drift velocity using the offline laser tracks calbration

 run       - run number
 timeStamp - tim stamp in seconds
 deltaT    - integration period to calculate time0 offset
 side      - 0 - A side,  1 - C side, 2 - mean from both sides
 Note in case no data form both A and C side - the value from active side used

 Get the correction of the drift velocity using the online laser tracks calbration

 run       - run number
 timeStamp - tim stamp in seconds
 deltaT    - integration period to calculate time0 offset
 side      - 0 - A side,  1 - C side, 2 - mean from both sides
 Note in case no data form both A and C side - the value from active side used

 common drift velocity retrieval for online and offline method

 Get the correction of the drift velocity using the CE laser data
 combining information from the CE,  laser track calibration
 and P/T calibration

 run       - run number
 timeStamp - tim stamp in seconds
 deltaT    - integration period to calculate time0 offset
 side      - 0 - A side,  1 - C side, 2 - mean from both sides

 return drift velocity using the TPC-ITS matchin method
 return also distance to the closest point

 Get time dependent time 0 (trigger delay in cm) correction
 Arguments:
 timestamp - timestamp
 run       - run number

 Notice - Extrapolation outside of calibration range  - using constant function

 VERY obscure method - we need something in framework
 Find the TPC runs with temperature OCDB entry
 cache the start and end of the run

 binary search - find the run for given time stamp

 filter outlyer measurement
 Only points around median +- sigmaCut filtered

 filter outlyer measurement
 Only points around median +- cut filtered

 filter outlyer measurement
 Only points with normalized errors median +- sigmaCut filtered

 sort array - neccessay for approx

 smmoth graph - mean on the interval

 Use constant interpolation outside of range

 Use constant interpolation outside of range also for spline fits

 Filter DCS sensor information
   ymin     - minimal value
   ymax     - max value
   maxdy    - maximal deirivative
   sigmaCut - cut on values and derivative in terms of RMS distribution
 Return value - accepted fraction

 Algorithm:

 0. Calculate median and rms of values in specified range
 1. Filter out outliers - median+-sigmaCut*rms
    values replaced by median

 Filter temperature array
 tempArray    - array of temperatures         -
 ymin         - minimal accepted temperature  - default 15
 ymax         - maximal accepted temperature  - default 22
 sigmaCut     - values filtered on interval median+-sigmaCut*rms - defaut 5
 return value - fraction of filtered sensors

 Filter CE data
 Input parameters:
    deltaT   - smoothing window (in seconds)
    cutAbs   - max distance of the time info to the median (in time bins)
    cutSigma - max distance (in the RMS)
    pcstream - optional debug streamer to store original and filtered info
 Hardwired parameters:
    kMinPoints =10;       // minimal number of points to define the CE
    kMinSectors=12;       // minimal number of sectors to define sideCE
 Algorithm:
 0. Filter almost emty graphs (kMinPoints=10)
 1. calculate median and RMS per side
 2. Filter graphs - in respect with side medians
                  - cutAbs and cutDelta used
 3. Cut in respect wit the graph median - cutAbs and cutRMS used
 4. Calculate mean for A side and C side

 Filter Drift velocity measurement using the tracks
 0.  remove outlyers - error based
     cutSigma

 get laser time offset
 median around timeStamp+-deltaT
 QA - chi2 needed for later usage - to be added
    - currently cut on error

 Filter Goofie data
 goofieArray - points will be filtered
 deltaT      - smmothing time window
 cutSigma    - outler sigma cut in rms
 minVn, maxVd- range absolute cut for variable vd/pt
             - to be tuned

 Ignore goofie if not enough points

 Make a statistic matrix
 Input parameters:
   array        - TObjArray of AliRelKalmanAlign
   minFraction  - minimal ration of accepted tracks
   minStat      - minimal statistic (number of accepted tracks)
   maxvd        - maximal deviation for the 1
 Output matrix:
    columns    - Mean, Median, RMS
    row        - parameter type (rotation[3], translation[3], drift[3])

 Smooth the array of AliRelKalmanAlign - detector alignment and drift calibration)
 Input:
   array     - input array
   stat      - mean parameters statistic
   direction -
   sigmaCut  - maximal allowed deviation from mean in terms of RMS

 Merge 2 RelKalman arrays
 Input:
   arrayP    - rel kalman in direction plus
   arrayM    - rel kalman in direction minus

 baseDir:   Base directory with the raw Kr calibration trees
            and the trees from the calibQA
            it assumes to following structure below:
            KryptonCalib/<year>/calibKr/calibKr.<year>.<id>.root
            calibQAdEdx/<year>/calibQA.<year>.<perid>.tree.root
            map/treeMapping.root

 baseDir:   Base directory with Pulser information
 TTrees are added to the base tree as a friend tree

 === add the calibPulser trees ======================================

 baseDir:   Base directory with Distortion information
 TTrees are added to the base tree as a friend tree
 If base tree not provide - first tree from list is used as base

 === add the calibDistortion trees ======================================
TString inputTreesDistortionCalib       = gSystem->GetFromPipe(Form("ls %s/calibDistortion/20*/*.tree.root",baseDir.Data()));
 TString baseDir="$NOTES/reconstruction/distortionFit/"; TTree *tMain=0;
 AliTPCcalibDButil::ConnectDistortionTrees("$NOTES/reconstruction/distortionFit/", "calibTimeResHisto.root", 0);

 baseDir:   Base directory with per Pad information
 TTrees are added to the base tree as a friend tree
 Example usage
   TString baseDir="/hera/alice/fsozzi/summarymaps/calib2/";  // prefix directory with calibration with slash at the end
   TString pattern="QA/*/*root";
   TTree * tree =  AliTPCcalibDButil::ConnectCalPadTrees(baseDir,pattern,0);   //create tree and attach calibration as friends

getter preprocess information

getters ref data

setters for pad by pad information