Support `@task.spark_submit` or `@task.pyspark_submit`

### Description Hello, I'm an user of Airflow for managing Spark jobs. I was wondering if there's any plan to provide `@task.spark_submit` decorator in the future. Actually, I’ve been using a custom provider that I made myself. In my situation, I write and execute pyspark code locally, and thenI transform the code to a form that can be submitted to YARN finally. Transforming the code into a form suitable for YARN (just creating separate .py file) is not particularly difficult. However, when the task code and the YARN submission code are separated, the folder structure within the DAG can become complex, and the task code and PySpark job files cannot reference each other in the IDE. by `@task.spark_submit` (or `@task.pyspark_submit`?), it would improve development productivity significantly and very convenient. This is because we could replace the existing @task.pyspark decorators by simply changing the decorator name and injecting the YARN options. > all I need to do is just changing the decorator name and inject the YARN options. ### Use case/motivation The idea is to borrow the approach from the existing `@task.pyspark` and `@task.docker` decorators. It works by writing a temporary py file to /tmp and then submitting that file. ```python class _SparkSubmitDecoratedOperator(DecoratedOperator, SparkSubmitOperator): custom_operator_name = "@task.spark_submit" # In the constructor, Some codes like removing the sc and spark arguments are located . ... def execute(self, context: Context): with TemporaryDirectory() as tmp_dir: input_filename = os.path.join(tmp_dir, "script.in") script_filename = os.path.join(tmp_dir, "script.py") output_filename = os.path.join(tmp_dir, "script.out") error_filename = os.path.join(tmp_dir, "script.err") with open(input_filename, "w", encoding="utf-8") as file: if self.op_args or self.op_kwargs: self.pickling_li…