[FEA] Qualification should treat json_tuple function as supported

[cindyyuanjiang]

The json_tuple function is implemented in the rapids-plugin side. We need to update this in the tools repo.

Generate a CPU eventlog by running below query in spark-sql:

SELECT json_tuple('{"a":1, "b":2}', 'a', 'b');