merge_models does not merge list even if it does not include Link

[tomohirohiratsuka]

Hi, I'm wondering this is expected behavior or not.
When I update document with update method, it returns updated document only when the updated property is not list.
I found that inside update method it merging via parsing.merge_models.
merge_models looks like care about if the list includes Link or not but it does not merge even when no Links are found in the list.
Is this expected behavior?
I'd appreciate it, if anyone answer.
Thank you.

@wrap_with_actions(EventTypes.UPDATE)
    @save_state_after
    async def update(
        self,
        *args,
        ignore_revision: bool = False,
        session: Optional[ClientSession] = None,
        bulk_writer: Optional[BulkWriter] = None,
        skip_actions: Optional[List[Union[ActionDirections, str]]] = None,
        skip_sync: Optional[bool] = None,
        **pymongo_kwargs,
    ) -> DocType:
        """
        Partially update the document in the database

        :param args: *Union[dict, Mapping] - the modifications to apply.
        :param session: ClientSession - pymongo session.
        :param ignore_revision: bool - force update. Will update even if revision id is not the same, as stored
        :param bulk_writer: "BulkWriter" - Beanie bulk writer
        :param pymongo_kwargs: pymongo native parameters for update operation
        :return: None
        """

        arguments = list(args)

        if skip_sync is not None:
            raise DeprecationWarning(
                "skip_sync parameter is not supported. The document get synced always using atomic operation."
            )
        use_revision_id = self.get_settings().use_revision

        if self.id is not None:
            find_query: Dict[str, Any] = {"_id": self.id}
        else:
            find_query = {"_id": PydanticObjectId()}

        if use_revision_id and not ignore_revision:
            find_query["revision_id"] = self._previous_revision_id

        if use_revision_id:
            arguments.append(SetRevisionId(self.revision_id))
        try:
            result = await self.find_one(find_query).update(
                *arguments,
                session=session,
                response_type=UpdateResponse.NEW_DOCUMENT,
                bulk_writer=bulk_writer,
                **pymongo_kwargs,
            )
        except DuplicateKeyError:
            raise RevisionIdWasChanged
        if bulk_writer is None:
            if use_revision_id and not ignore_revision and result is None:
                raise RevisionIdWasChanged
            merge_models(self, result) # this update self
        return self

def merge_models(left: BaseModel, right: BaseModel) -> None:
    from beanie.odm.fields import Link

    if hasattr(left, "_previous_revision_id") and hasattr(
        right, "_previous_revision_id"
    ):
        left._previous_revision_id = right._previous_revision_id  # type: ignore
    for k, right_value in right.__iter__():
        left_value = left.__getattribute__(k)
        if isinstance(right_value, BaseModel) and isinstance(
            left_value, BaseModel
        ):
            merge_models(left_value, right_value)
            continue
        if isinstance(right_value, list):
            links_found = False
            for i in right_value:
                if isinstance(i, Link):
                    links_found = True
                    break
            if links_found: # why no links case is not handled?
                continue
        elif not isinstance(right_value, Link):
            left.__setattr__(k, right_value)

[roman-right]

Hi! I'll check it out on the next bug-fixing session this or next week

[roman-right]

Thank you for the catch

[tomohirohiratsuka]

@roman-right
Hi, Thank you for your response.
I've tried to fix this with this PR
If I did something wrong, please tell me.
Thank you.