class HKalFilterTask: public HReconstructor

 Create a new kalsys object depending on the class variables
 bWire, bDaf and bCompHits.
 An existing kalsys object will be deleted.

 Fills categories with reconstruction results.
 pMetaMatch: Meta match candidate. Index to Kalman filter is set.
 allhitsGeantMdc:

 Fills categories for measurement sites and hits with reconstruction results.
 Storing these categories consumes large amounts of disk space and is mainly
 intended for debugging purposes.

 Returns the indices for the first and last measurement site needed by the
 main track category.

 output:
 iFirstSite: index of first measurement site in catagory catKalSite.
 iLastSite:  index of last measurement site in catagory catKalSite.

 input:
 allhitsGeantMdc: Array with Geant hits in MDCs.

 Task to run the Kalman filter.

 Input:
 ref:   Run ID number.
 sim:   Simulation/Data.
 wire:  Run Kalman filter Mdc drift chamber hits or on segment hits.
 daf:   Apply the Deterministic Annealing Filter (DAF).

 Get particle hypothesis.

 Calculates initial state vector and covariance matrix for
 the Kalman filter.
 Position is derived from first measurement,
 direction from MDC segments and momentum is taken from Spline fit.

 output:
 inisv: Initial state vector.
 iniC:  Initial covariance matrix.
 input:
 pid:        GEANT particle ID.
 allhits:    Array with MDC hits. Must be of class HKalMdcHit.
 pMetaMatch: Pointer to Meta match.
 inSeg:      Pointer to inner MDC segment.
 outSeg:     Pointer to outer MDC segment.

 Calculate starting position and direction of Geant hit.

 Retrieve initial state from Runge-Kutta fit.
 Find any Runge-Kutta object from MetaMatch.

 Returns true if particles with Geant ID pid are going to be filtered.

 Fill the array allhits with the next track candidate.

 Error estimates used to fill the initial covariance matrix.
 dx, dy:   Position errors in sector coordinates and mm.
 dtx, dty: tx = tan(px/pz), ty = tan(py/pz) in radians.
 dmom:     Momentum uncertainty as a fraction. For example 0.1 means initial estimate is
           within 10% of the true momentum.

 Only reconstruct particles with certain pid.
 pid: Array containing the Geant particle ids.
 n:   Size of array pid. Pass n=0 to reconstruct all particles.

 Modifies the type of the Kalman filter that will be applied:
 wire:  Run Kalman filter Mdc drift chamber hits or on segment hits.
 daf:   Apply the Deterministic Annealing Filter (DAF).
 comp:  Consider competing hits for the DAF.

 It is recommended to specify the desired values for these parameters
 already in the contructor.
 Returns kFALSE if the parameter combination is not valid.
 An existing Kalman filter object will be deleted along with its
 settings.

 Set verbose level:
 1: print warnings and error messages
 2: print error messages only
 all else: print nothing.