working apply functionality

[mikeAdamss]

allow passing of functions into a databaker dimension constructor to allow any required override/modification of dimension headers before they are pivoted.

HDim(cells, "My Dim", CLOSEST, ABOVE, apply=my_function)

or even

HDim(cells, "My Dim", CLOSEST, ABOVE, apply=lambda x: x.strip())

So basically a databaker parallel of the pandas apply() method. Is uber performant (changing a single string beats changing a subset of a series hands down) and should lead to cleaner code with much less of a reliance on post processing.

also test feature for same plus corrected a given to a then for test readability in an unconnected scenario (spotted in passing).

[mikeAdamss]

need to extend this to optionally take a tuple of functions, eg:

HDim(cells, "My Dim", CLOSEST, ABOVE, apply=(my_function, lambda x: (int(x)))

as there's bound to be times we want to make more than one intervention at the HDim constructor level.