Conserve python object in callbacks

_Issue_
Consider this code
```python
class MyObj_Is_An_Array:
def __init__():
self.a_data_to_access = 42
def __descriptor__(): # Give DaCe the possibility to use it as an array
...

# For the sake of the argument this function will _not_ be parsed by DaCe
# but triggers a callback
def callback(the_obj):
the_obj.a_data_to_access # <--- error, we get a dace.Array instead of a MyObj_Is_An_Array

@dace.program
def a_program(the_obj):
...
callback(the_obj)

the_obj = MyObj_Is_An_Array()
a_program(the_obj)
```

_Solution_
Marshall the true type of Python object in the callback system.

(Also we would need it to keep working with literal list, e.g. `callback([obj, obj, obj])`

_Motivation_

The [Pace code](https://github.com/NOAA-GFDL/pace) makes use of callback to access an optimized halo exchange system. Of course we should ultimately try and have the halo exchange be DaCe readable, but in the meantime callbacks are an effective (and used) way to reach working system in large codebase.