High-level access for deep copy

[drahnreb]

I used to do this (deprecated) for a deep copy:

>>> import awkward1 as ak
>>> array = ak.Array([])
>>> array.copy() # deprecated with removal of pandas integration #350 

You could do this:

>>> ak.Array(array.layout.deep_copy()) # unwrap and rewrap required similar to #496 
>>> # or
>>> import copy
>>> copy.deepcopy(array)

Would be nice to have something like this (with the same kwargs as the low-level?):

>>> ak.copy(array) # or registered np.copy() which would be strictly deep as object elements aren't supported by awkward1

[jpivarski]

A low level function exists for this, so it will be a matter of connecting it to the high level. Overriding __copy__ and __deepcopy__ are good ideas, too.