Nicer default printing for estimators and other types

[kahaaga]

Estimators and other types need nicer default printing. Most of our types are parameterized, often with multiple type parameters, leading to extremely long and messy printing in the console. It would be nice to clean this up a bit.

Example:

julia> s = StatisticalComplexity()
StatisticalComplexity{MLE{SymbolicPermutation{3, typeof(ComplexityMeasures.isless_rand)}}, Distances.JSDivergence, Renyi{Float64, Float64}}(Distances.JSDivergence(), MLE{SymbolicPermutation{3, typeof(ComplexityMeasures.isless_rand)}}(SymbolicPermutation{3, typeof(ComplexityMeasures.isless_rand)}(OrdinalPatternEncoding{3, typeof(ComplexityMeasures.isless_rand)}([0, 0, 0], ComplexityMeasures.isless_rand), 1)), Renyi{Float64, Float64}(1.0, 2.0), Base.RefValue{Float64}(0.0))

It would be nice if this just printed something like

julia> s = StatisticalComplexity()
StatisticalComplexity(probest=MLE{SymbolicPermutation}, dist=Distances.JSDivergence, entr = Renyi)

I'm not sure what level of abstraction default printing should have. But I think a clean-up, all the way from the information measure definitions, all the way up to nested types such as StatisticalComplexity is in order.

[Datseris]

we could define a show method for all top level abstract types that shows the fields and values. We can make the arbitrary decision that for some types we print the actual types, while for all other types we print the type name.

X with fields:
  a = 0.5 # print value
  b = Renyi  # print type, $(nameof(typeof(value))

[kahaaga]

we could define a show method for all top level abstract types that shows the fields and values.

I guess using fieldnames(x::AbstractType) could be used to iterate through all fields and show them.