[SPARK-56612][PYTHON] Unify verify_result and container-type checks into verify_return_type helper

[Yicong-Huang]

What changes were proposed in this pull request?

Introduce verify_return_type(result, expected_type) in worker.py:

• concrete type (e.g. pa.Table) — isinstance check, return result.
• Iterator[T] — check result is an Iterator, return a lazy iterator that checks each element against T on consumption.

Apply it to:

• SQL_MAP_ARROW_ITER_UDF and SQL_SCALAR_ARROW_ITER_UDF (replacing the old verify_result factory).
• verify_arrow_table / verify_arrow_batch — inlined at their 3 call sites and removed.

Runtime now strictly requires an Iterator (was: any iterable), matching the declared contract ArrowMapIterFunction = Callable[[Iterator[...]], Iterator[...]] and the _name == "Iterator" convention in pyspark/sql/pandas/typehints.py.

Why are the changes needed?

Part of SPARK-55388. One well-named helper unifies the iterator factory and scattered inline isinstance checks, enabling further eval-type refactors.

Does this PR introduce any user-facing change?

No. Minor tightening: a UDF returning a non-Iterator iterable (e.g. list) is now rejected with UDF_RETURN_TYPE, aligning runtime with the documented Iterator[...] signatures.

How was this patch tested?

Existing integration coverage: test_arrow_map.py::test_other_than_recordbatch_iter exercises the iterator path; other verify_arrow_* callers (mapInArrow, grouped-map Arrow UDFs) exercise the concrete-type path.

Was this patch authored or co-authored using generative AI tooling?

No

[Yicong-Huang]

@gaogaotiantian please kindly share your thoughts!

cc @zhengruifeng @HyukjinKwon