generic Transform(json=) instantiation

[djkapner]

It would be nice to be able to load a transform from a dict without knowing its exact class ahead of time. But, not all the methods come through. Needs something fancier for it to become the correct object:

> r = renderapi.transform.AffineModel()
> np.array([name for name in dir(r) if callable(getattr(r, name))])
array(['__class__', '__delattr__', '__dir__', '__eq__', '__format__',
       '__ge__', '__getattribute__', '__gt__', '__hash__', '__init__',
       '__init_subclass__', '__le__', '__lt__', '__ne__', '__new__',
       '__reduce__', '__reduce_ex__', '__repr__', '__setattr__',
       '__sizeof__', '__str__', '__subclasshook__', '_process_dataString',
       'calc_properties', 'concatenate', 'convert_points_vector_to_array',
       'convert_to_point_vector', 'estimate', 'fit', 'from_dict',
       'inverse_tform', 'invert', 'load_M', 'tform', 'to_dict'],
      dtype='<U30')

but:

> f = renderapi.transform.Transform(json=r.to_dict())
> np.array([name for name in dir(f) if callable(getattr(f, name))])
array(['__class__', '__delattr__', '__dir__', '__eq__', '__format__',
       '__ge__', '__getattribute__', '__gt__', '__hash__', '__init__',
       '__init_subclass__', '__le__', '__lt__', '__ne__', '__new__',
       '__reduce__', '__reduce_ex__', '__repr__', '__setattr__',
       '__sizeof__', '__str__', '__subclasshook__', '_process_dataString',
       'from_dict', 'to_dict'], dtype='<U19')

[djkapner]

I see renderapi.transform.load_transform_json() does this