v1.4.6.0

v1.4.6.0

New features:

1. Add new method dct_params for ROOT.RooFitResult, that gets a dictionary of all parameter values
2. plotting/canvas.py: Add option to save the canvas congent in tar/tgz/zip-archives:

canvas >> 'test.zip'
canvas >> 'test.tar'
canvas >> 'test.tgz'

1. ostap/fitting/toys.py new module with simple functions to perform fitting toys

pdf = ...
results , stats = make_toys ( pdf      ,           ## PDF  to use 
                  1000                 ,           ## number of toys 
                  [ 'mass' ]           ,           ## varibales in dataset 
                  { 'nEvents' : 5000 } ,           ## configuration of pdf.generate
                  { 'ncpus'   : 2    } ,           ## configuration of pdf.fitTo
                  { 'mean' : 0.0 , 'sigma' : 1.0 } ## parameters to use for generation 
                 )

1. ostap/parallel/parallel_toys.py a version of fnuctoion above for the
   parallel execution via multiprocessing and/or parallel python)
2. add method evaluate for ROOT.RooFitResult to evaluate the arbintrary function of
   fit-parameters with uncertaionties

res = ... ## ROOT.RooFitResult object
fun = lambda x,y,z :  x*x+y*y+z*z 
val = res.evaluate ( fun , ('x' , 'y' , 'z') ) 

1. add EvalNVEcov and EvalNVEcor evaluators into ostap.math.derivative module.
   These objects evalaute the value of function of N-argument taking into account
   the uncertainties and correlations
2. add method max_cor to ROOT.RooFitResult to obtain the maximal correlation coefficient

res = ...  ## ROOT.RooFitResult
coefficient , variable = res.max_cor ( 'X' ) ## get the maximal correlation

1. add the column Global/max correlation for the table form of ROOT.RooFitResult
2. add ncpus argument to ostap.parallel.parallel_toys.parallel_toys

Backward incompatible changes

Bug fixes:

1. Minor fix in Files.hadd
2. Couple of minor fixes in
   • ROOT.RooFitResult.sum
   • ROOT.RooFitResult.multiply
3. Couple of minor fixes in ostap.math.derivative