Hi , I'm having a hard time writing my custom processor using python extension. it seems this feature is not mature as I thought or maybe I'm miss understanding something. To summarize I have encountered two issues but they are both related to the same error, as follows: 1- I was experimenting with developing a custom processor using the RecordTransform class to partition a json array, my  processor is simple, it takes json array of records where each record is a json object that has a "name" and "age" fields. The processor then try to return RecordTransformResult from the transform method as follows:   def transform(self, context, record, schema, attributemap): return RecordTransformResult(schema=None,record=record,relationship="success",partition={'name':record['name']})   I'm trying to use partition to split the record and passing the partition key just as described in the developer guide: https://nifi.apache.org/documentation/nifi-2.0.0-M1/html/python-developer-guide.html#record-transform When I place the processor in the extensions folder and run nifi, Im getting the following error in the nifi-python.log:   AttributeError: 'dict' object has no attribute '_get_object_id'   After a lot time researching the issue I found that the py4j expects a java type dictionary and a conversion should happen  based on the post: https://stackoverflow.com/questions/57502975/py4j-serialization-attributeerror-dict-object-has-no-attribute-get-object I was not sure how to access the java gateway object from the processor but after looking at the code from the framework folder I found in the controller.py how to instantiate it so I copied the same code and my transform method became like this:   def transform(self, context, record, schema, attributemap): java_port = int(os.getenv('JAVA_PORT')) auth_token = os.getenv('AUTH_TOKEN') gateway = JavaGateway( callback_server_parameters=CallbackServerParameters(port=0, auth_token=auth_token), gateway_parameters=GatewayParameters(port=java_port, read_timeout=None, enable_memory_management=True, auth_token=auth_token), auto_convert=True) input_dict = {'name':record['name']} mc_run_map_dict = MapConverter().convert(input_dict,gateway._gateway_client) return RecordTransformResult(schema=None,record=record,relationship="success",partition=mc_run_map_dict)   The code above worked and the record was split accordingly. My question is this the correct way or am I missing something because I thought I followed the steps in the developer guide? 2- The other scenario is when I tried to override the getRelationships to define my custom ones again as described in the developer guide, as follows :   def getRelationships(self): failedrel = Relationship(name="MyFailedRel",description="custom failed rel") successrel = Relationship(name="success",description="custom success rel") return [failedrel,successrel]   First I got the error that List doesnt have the attribute "_get_object_id" , so I tried to covert the list using the method above but now I'm getting the following error:   AttributeError: 'Relationship' object has no attribute '_get_object_id'   Im not sure what to do about this.  It could be related to the  first scenario. Any suggestion would be really appreciated as I have been struggling with this for couple of days. @MattWho  , @cotopaul , @steven-matison , @joseomjr  Thanks   ... View more