get_scenes() fails with invalid type error

[BenoitGodin]

Whenver I call get_scenes, I get the following error

import dirigera
dirigera_hub = dirigera.Hub(token="XXXXXXXXX", ip_address="XXXXXXXXX")
dirigera_hub.get_scenes()
Traceback (most recent call last):
File "", line 1, in
File "/usr/local/lib/python3.10/dist-packages/dirigera/hub/hub.py", line 347, in get_scenes
return [dict_to_scene(scene, self) for scene in scenes]
File "/usr/local/lib/python3.10/dist-packages/dirigera/hub/hub.py", line 347, in
return [dict_to_scene(scene, self) for scene in scenes]
File "/usr/local/lib/python3.10/dist-packages/dirigera/devices/scene.py", line 148, in dict_to_scene
return Scene(dirigeraClient=dirigera_client, **data)
File "/usr/local/lib/python3.10/dist-packages/pydantic/main.py", line 253, in init
validated_self = self.pydantic_validator.validate_python(data, self_instance=self)
pydantic_core._pydantic_core.ValidationError: 1 validation error for Scene
commands.0
Input should be a valid string [type=string_type, input_value={'type': 'device', 'id': ...e_open', 'volume': 93}]}, input_type=dict]
For further information visit https://errors.pydantic.dev/2.11/v/string_type

It seems attributes of one of the device is not correctly parsed.

According to input_type value, this is about an open/close sensor for which I configured to play a chime (volume 93) when the state changes to open. Can you have a look at this?

[BenoitGodin]

Digging further, I noticed get_scenes() fails on an the first scene, that is part of the an open_close device

It's not a scene that I created, but a scene related to the configuration of the device (play a sound when the state changes to open)
Here's the dict (data) provide to the creation of a Scene object is:
{'id': '8ed0535a-c2a6-449b-a3f5-8c2f7f1a4b1b', 'info': {'name': 'open scene', 'icon': 'scenes_clean_sparkles'}, 'type': 'customScene', 'triggers': [{'id': 'e0fce548-ca85-4087-aff0-e50499193b11', 'type': 'app', 'disabled': False}, {'id': '429d0ffb-0d05-4478-87c8-170681d4e73a', 'type': 'event', 'disabled': False, 'trigger': {'days': ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun'], 'deviceId': '67d26ce4-99f0-4571-ba39-96747b8ab763_1', 'filter': {'attribute': {'isOpen': True}}, 'notifications': [{'type': 'opened'}]}}, {'id': '77882143-f692-44fd-bd9a-85a411355d52', 'type': 'event', 'disabled': False, 'trigger': {'days': ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun'], 'undo': True, 'deviceId': '67d26ce4-99f0-4571-ba39-96747b8ab763_1', 'filter': {'attribute': {'isOpen': False}}}}], 'actions': [], 'commands': [{'type': 'device', 'id': '992002e1-6cbf-470d-b308-8787d0b8e1a8_1', 'enabled': True, 'commands': [{'type': 'playAudioClipByName', 'playAudioClipName': 'ikea://notify/03_device_open', 'volume': 74}]}], 'createdAt': '2025-07-31T16:03:54.070Z', 'lastCompleted': '2025-12-19T08:05:30.265Z', 'lastTriggered': '2025-12-19T08:05:30.265Z', 'undoAllowedDuration': 1, 'lastUndo': '2025-12-19T08:05:32.990Z'}

and the code fails on:

'commands': [
   {
      'type': 'device',
      'id': '992002e1-6cbf-470d-b308-8787d0b8e1a8_1',
      'enabled': True,
      'commands': [
         {
            'type': 'playAudioClipByName',
            'playAudioClipName': 'ikea://notify/03_device_open',
            'volume': 74
         }
      ]
   }
]

Skipping this scene, with is ugly change:
line 347 in hub.py:
347 return [dict_to_scene(scene, self) for scene in scenes[1:]]

makes the execution of get_scenes() successful

I think that somehow, the scenes not created by the end user should be discarded

[BenoitGodin]

And a correct solution to the above isse is to update the definition of the class Scene like this:

class Scene(BaseIkeaModel):
    dirigera_client: AbstractSmartHomeHub
    id: str
    type: SceneType
    info: Info
    triggers: List[Trigger]
    actions: List[Action]
    created_at: datetime.datetime
    last_completed: Optional[datetime.datetime] = None
    last_triggered: Optional[datetime.datetime] = None
    last_undo: Optional[datetime.datetime] = None
    #commands: List[str]
    commands: List[Any]
    undo_allowed_duration: int

(file scene.py, lines 122 - 134)

So the commands attribute can be a list of strings or a list of dicts