[SPARK-56463][PYTHON] Disallow unpickling UDT

### What changes were proposed in this pull request?

Disallow unpickling UDT.

### Why are the changes needed?

### Does this PR introduce _any_ user-facing change?

Yes, unpickling UDT is not supported.

### How was this patch tested?

CI.

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

No.

Closes #55330 from gaogaotiantian/disallow-unpickling.

Authored-by: Tian Gao <gaogaotiantian@hotmail.com>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>

Author: gaogaotiantian

python/pyspark/sql/types.py

@@ -1725,8 +1725,10 @@ def fromJson(cls, json: Dict[str, Any]) -> "UserDefinedType":
         pyClass = pyUDT[split + 1 :]
         m = __import__(pyModule, globals(), locals(), [pyClass])
         if not hasattr(m, pyClass):
-            s = base64.b64decode(json["serializedClass"].encode("utf-8"))
-            UDT = CloudPickleSerializer().loads(s)
+            raise PySparkValueError(
+                errorClass="UNSUPPORTED_OPERATION",
+                messageParameters={"operation": "unpickling user defined types"},
+            )
         else:
             UDT = getattr(m, pyClass)
         return UDT()

python/pyspark/testing/objects.py

@@ -90,7 +90,7 @@ def sqlType(cls):
 
     @classmethod
     def module(cls):
-        return "__main__"
+        return "pyspark.testing.objects"
 
     def serialize(self, obj):
         return [obj.x, obj.y]