Removing feature: send context in venv operators (using use_airflow_context)

[amoghrajesh]

related: #45739

#41363 is a feature that heavily relied on pydantic models of AIP 44 to do some serialisation magic.

Historically did this:

            if self.use_airflow_context:
                from airflow.serialization.serialized_objects import BaseSerialization

                context = get_current_context()
                with create_session() as session:
                    # FIXME: DetachedInstanceError
                    dag_run, task_instance = context["dag_run"], context["task_instance"]
                    session.add_all([dag_run, task_instance])
                    serializable_context: dict[Encoding, Any] = BaseSerialization.serialize(
                        context, use_pydantic_models=True
                    )

• So it used a pydantic model to serialize the DAG.
• And the mapping looked like so:

_type_to_class: dict[DAT | str, list] = {
    DAT.BASE_JOB: [JobPydantic, Job],
    DAT.TASK_INSTANCE: [TaskInstancePydantic, TaskInstance],
    DAT.DAG_RUN: [DagRunPydantic, DagRun],
    DAT.DAG_MODEL: [DagModelPydantic, DagModel],
    DAT.LOG_TEMPLATE: [LogTemplatePydantic, LogTemplate],
    DAT.DATA_SET: [DatasetPydantic, Dataset],
    DAT.TRIGGER: [TriggerPydantic, Trigger],
}

• And this:

def serialize_operator(x: DAG) -> dict:
    from airflow.serialization.serialized_objects import SerializedDAG

    return SerializedDAG.serialize_dag(x)


def validate_operator(x: DAG | dict[str, Any], _info: ValidationInfo) -> Any:
    from airflow.serialization.serialized_objects import SerializedDAG

    if isinstance(x, DAG):
        return x
    return SerializedDAG.deserialize_dag(x)

Since now we do not have the pydantic models of AIP 44 AND the fact that this didnt make it to any release, now is a time to chop it off.

cc: @phi-friday

---

^ Add meaningful description above
Read the Pull Request Guidelines for more information.
In case of fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in a newsfragment file, named {pr_number}.significant.rst or {issue_number}.significant.rst, in newsfragments.

[ashb]

For context on why we are removing this:

1. It never actually made it in to a release (it was tagged with 2.10.0 milestone but somehow got missed in that release)
2. It relies on AIP-44 and the pydantic models it creates to work, and those have been removed in main/Airflow 3
3. With the split out of the new standard provider it will be easier to add this back in quicker, but trying to support it now will delay Airflow 3 work as we have a long list of critical features related ti AIP-72/Task Execution interface, and unfortunately we have to cut something.

Sorry to revert this yet again :(

[amoghrajesh]

@ephraimbuddy are we good to send this one in?

[potiuk]

Makes sense. Thanks for the reasoning @ashb . Should we also add an issue for 3.1 to re-add it ?

[amoghrajesh]

Makes sense. Thanks for the reasoning @ashb . Should we also add an issue for 3.1 to re-add it ?

Thanks @potiuk, created #46361 for tracking sake.