Make param validation consistent for DAG validation and triggering

[jscheffl]

This PR adjusts the DAG level Params validation to the validation also used during triggering and DAG constructor.

It resolves the problem reported in #34227 that params with a value of None are leading to invalid/unparsable DAGs after upgrade to Airflow 2.7.1 as a side effect of #33141 - reason is that the validation previously checked for has_value assuming that the schema also defines a required field.

How to test:

• See all tests are green
• Define a DAG with the code below and see that with the fix applied, the DAG is parse-able again and can be used:

@dag(
    dag_id="dag-example",
    start_date=datetime(2023, 2, 2, tz="America/New_York"),
    schedule="41 2 * * *",
    params={
        "any_param": None,
    },
):
    ...

closes: #34227
related: #31299, #31301