Missing Phase Attributes in attr_list_vect

[DEUCE1957]

Environment

• Grid2op version: 1.11.0
• System: Arch Linux

Bug description

Expect to find the attributes "load_theta", "gen_theta", "theta_or", "theta_ex", "storage_theta" in CompleteObservation.attr_list_vect. However it is only in CompleteObservation.attr_list_json.

from grid2op.Observation import CompleteObservation
print("load_theta" in CompleteObservation.attr_list_vect) # False
print(load_theta" in CompleteObservation.attr_list_json) # True

Expected behaviour would be to have all attributes in attr_list_vect, including the phase attributes.

I wondered if it might have to do with the to_json() conversion, but internally they seem to be concatenated. From GridObjects._convert_to_json(...):

  @classmethod
  def _convert_to_json(cls, dict_: Dict[str, Any]) -> None:
      for attr_nm in cls.attr_list_vect + cls.attr_list_json: # Lists are concatenated
          tmp = dict_[attr_nm]
          dtype = tmp.dtype
          if dtype == dt_float:
              dict_[attr_nm] = [float(el) for el in tmp]
          elif dtype == dt_int:
              dict_[attr_nm] = [int(el) for el in tmp]
          elif dtype == dt_bool:
              dict_[attr_nm] = [bool(el) for el in tmp]
          elif dtype == float:
              dict_[attr_nm] = [float(el) for el in tmp]
          elif dtype == int:
              dict_[attr_nm] = [int(el) for el in tmp]
          elif dtype == bool:
              dict_[attr_nm] = [bool(el) for el in tmp]

In the docstring of the method GridObjects.to_json(), we get a partial explanation:

Concretely, if `obs` is an observation (:class:`grid2op.Observation.BaseObservation`)
            then `obs.to_dict()` will have the keys `type(obs).attr_list_vect` and the values will
            be numpy arrays whereas `obs.to_json()` will have the keys
            `type(obs).attr_list_vect` and `type(obs).attr_list_json` and the values will be
            lists (serializable)

However, obs.load_theta, for instance is a numpy array already, so I don't understand why it cannot be included in to_dict() by default.

[DEUCE1957]

Ignore the bug label, as this is more of a question