[WIP] fix #86c, #37, #39, added support to many numpy functions

[impact27]

• added a asQuantity function to transform an array into a Quantity (a bit like numpy's asarray)
• Allowed transformation of array of Quantity to Quantity
• A dimensionless 0 should be compared and added (Remove the distinction between 0 and dimensionless zero)
• A Quantity is not an ndarray! is has an ndarray. Therefore it should not respond to ['__array_interface__', '__array_struct__' , '__array__', '_data']
• added __array_ufunc__ to handle numpy function

[impact27]

The fail with pandas comes from the fact that pandas can't really test for nan with the new implementation. I am not sure how to fix it.
The relevant code is in pandas/core/dtypes/missing.py

def _isna_new(obj):
    if is_scalar(obj):
        return libmissing.checknull(obj)
    # hack (for now) because MI registers as ndarray
    elif isinstance(obj, ABCMultiIndex):
        raise NotImplementedError("isna is not defined for MultiIndex")
    elif isinstance(obj, (ABCSeries, np.ndarray, ABCIndexClass,
                          ABCExtensionArray)):
        return _isna_ndarraylike(obj)
    elif isinstance(obj, ABCGeneric):
        return obj._constructor(obj._data.isna(func=isna))
    elif isinstance(obj, list):
        return _isna_ndarraylike(np.asarray(obj, dtype=object))
    elif hasattr(obj, '__array__'):
        return _isna_ndarraylike(np.asarray(obj))
    else:
        return obj is None

Also, older versions of numpy fails because they want Quantity to be an array. The problem is that if Quantity gives an array representation, the units are lost. So these changes can only be applied to newer versions of numpy. Maybe the solution is to only return the __array_ methods if the version of numpy is below some number?

[impact27]

fix #714, kind of fix #484

[hgrecco]

This has been superseed by #905 and related