When I try to assert a dataframe using the PySpark API, if a dataframe is none, I do not get the assertion error, but instead, the method returns false. Is it a bug, or should I handle my test verification differently?