Lambda yield from 1M run 10783.

    First test on Lambda reconstruction in preparation to pass 2 was completed.
    The results are presented in   fig.1 
    and in   fig.2 

    The test was done with loosing selection criteria, e.g. track sharing for
    several vee hyphotesis was allowed, maximum om dca - distance of closest
    approach in m1 spectrometer was choosen to be 0.7 cm etc.
    
    Conclusion.
   ------------
    
 1. We see clear mass peaks for all  4 types of Lambdas: vx, Ux, 1p, 2p.

 2. We see more entries for each type. 

 3. Lambda mass is at PDG value (see narrow bins histograms).

 4. We see good antilambda peaks in each channel.
 
 5. Soft cut selector works well for vx type. One can use in addition
    target cut and pid (rich, etrd and photon) information. 
    Maybe default v0 setup will be enough to run recon for 2p type.
    I should imply more cut for 1p lambdas.
 
 6. I don't see Lambda 1520 decay (Lambda with 2 pions) but we have not
    enough statistics at this mass.
    
     Perspectives.
   -----------------
     
 1.  I have now the information (Fig.1) which will help me to increase
     Ux signal. This processor was tuned using GE781 data at lower Xf.
     In Monte Carlo run I have about equal vx and Ux signals.
     
 2.  I have to build sigma zero mass. I lost gamma 4-vector in this run.
 
 3.  I shall run 1M test again soon.
   
 
     Warning.
   ------------

     I added several histograms to see a sistematics with GE781 data
     
     analysis. They are presented here in   fig.3 

     One can see 
     
  a)  correct lambda mass;
  
  b)  correct vee angle;
  
  c)  systematically lower by 0.3% proton momentum with track in vertex;

  d)  systematically higher by 3.2% proton momentum with track in m1m2;
  
  
     Maybe 0.3% comes from additional pt-kick correction in Soap, which
     
     must be introduced in GE781 as well. But what can we do with 3% for
     
     m1m2 tracks ?