Background:
My function extracts data from a CSV file and generates JSON using two proto files. I need to add another three fields that are not in the proto files and generate a final json . Later, this JSON is serialized and sent to RabbitMQ.

My code:

import argparse
import logging
from python import model_pb2,events_pb2
from google.protobuf.timestamp_pb2 import Timestamp
import time
import decimal
import csv
#from proto_definitions import create_answer_type
from pathlib import Path
from google.protobuf.json_format import MessageToDict, MessageToJson

#from  python.model_pb2 import AnswerUpdateRequest
from google.protobuf import timestamp_pb2
from google.protobuf import wrappers_pb2 as wrappers
from google.protobuf import json_format

#from proto import model
import pandas as pd
import boto3
import dotenv
import os
import json
import chardet
from datetime import date
from document_parser.pdf import PdfParser
from document_parser.params import Paths,Enviorment,rabbitmq
from document_parser.utils import Timer
from document_parser.question_extractor import QuestionExtractor, load_questions, Question, ValidTypes
import pika
import json
import base64
from document_parser.uploads_downloads import *
from document_parser.status_file import *
from google.protobuf.timestamp_pb2 import Timestamp
from google.protobuf.json_format import MessageToJson, MessageToDict
from python.events_pb2 import AnswerUpdateRequest
from python.model_pb2 import AnswerSource,Answer,AnswerFieldType,Entity
from google.protobuf.json_format import Parse
from google.protobuf.timestamp_pb2 import Timestamp
from google.protobuf.json_format import MessageToJson, MessageToDict
from python.events_pb2 import AnswerUpdateRequest
from python.model_pb2 import AnswerSource,Answer,AnswerFieldType,Entity
    
def generate_final_json(json_decoded_data, document_type, file_generated_time, parsed_pdf_file_path):
        csv_path = Paths.llm_prediction_dir.joinpath(f"{parsed_pdf_file_path.parent.name}_extracted.csv")
        print("csv_path is " + str(csv_path))
      
    
        answer_update_request = AnswerUpdateRequest()
        answer_update_request.entity.type = Entity.ORGANIZATION
        answer_update_request.entity.id = json_decoded_data['orgId']
    
        with open(csv_path, 'r') as csv_file:
            csv_reader = csv.DictReader(csv_file)
            for row in csv_reader:
                
                if row["answer"] != "No answer":
                 #   answer_type = create_answer_type(row["answer"])
    
                    answer = model_pb2.Answer()
                    answer.key = row["api_field_name"]
    
                    answer.source.type = model_pb2.AnswerSource.DOCUMENT
                    answer.source.id = f"{json_decoded_data['documentKey']} | {document_type}"
                    
    
                    # provided_at_datetime = datetime.fromisoformat(json_decoded_data['provided_at'])
                    # provided_at_timestamp = Timestamp()
                    # provided_at_timestamp.FromDatetime(provided_at_datetime)
                    # answer.provided_at.CopyFrom(provided_at_timestamp)
    
                    provided_at = Timestamp()
                    provided_at_datetime = datetime.fromisoformat(file_generated_time)
                    answer.provided_at.CopyFrom(provided_at)
    
                    received_at = Timestamp()
                    received_at.FromDatetime(datetime.now())
                    answer.received_at.CopyFrom(received_at)
    
                #    answer.type = answer_type
                    answer.value.text = row["answer"]
    
    
                    answer_update_request.answers.append(answer)
        final_json={}
        
        # Convert the AnswerUpdateRequest to a JSON string
        answer_update_request_json = json_format.MessageToJson(answer_update_request, including_default_value_fields=True)
        answer_update_request_dict = json.loads(answer_update_request_json)
    
        # Create the final JSON object
        final_json = {
            "answerUpdateRequest": answer_update_request_dict,
            "documentKey": json_decoded_data['documentKey'],
            "applicationId": json_decoded_data["applicationId"],
            "activityId": json_decoded_data["activityId"]
        }
    
        # Write the final JSON to a file
        final_json_generated = json.dumps(final_json, indent=2)
        output_json_path = Paths.llm_prediction_dir.joinpath(f"{parsed_pdf_file_path.parent.name}_with_labels.json")
        with open(output_json_path, 'w') as json_file:
            json_file.write(final_json_generated)
            print(f"Protobuf message written to {output_json_path}")
    
        return // send serilaized message

as you can see below 3 keys are not there in proto files so I am manually adding them
“documentKey”: json_decoded_data[‘documentKey’],
“applicationId”: json_decoded_data[“applicationId”],
“activityId”: json_decoded_data[“activityId”]

model.proto:

syntax = "proto3";

package com.embroker.oracle;

import "google/protobuf/timestamp.proto";

option go_package = "github.com/embroker/oracle/sdk/go_sdk";

message Entity {
  Type type = 1;
  string id = 2;

  enum Type {
    ORGANIZATION = 0;
    USER = 1;
    APPLICATION = 2;
  }
}

message AnswerSource {
  Type type = 1;
  string id = 2;

  enum Type {
    UNKNOWN = 0;
    USER = 1;
    DOCUMENT = 2;
    EXTERNAL = 3;
  }
}

message Answer {
  string key = 1;
  AnswerSource source = 2;
  google.protobuf.Timestamp provided_at = 3;
  google.protobuf.Timestamp received_at = 4;
  AnswerFieldType type = 5;
  Value value = 6;

  message Value {
    oneof value {
      string text = 1;
      float decimal = 2;
      // ...
    }
  }
}

enum AnswerFieldType {
  ANSWER_FIELD_TYPE_UNSTRUCTURED = 0; // Can be useful for LLM purposes
  ANSWER_FIELD_TYPE_TEXT = 1;
  ANSWER_FIELD_TYPE_INTEGER = 2;
  ANSWER_FIELD_TYPE_BOOLEAN = 3;
  ANSWER_FIELD_TYPE_DECIMAL = 4;
  ANSWER_FIELD_TYPE_DATE = 5;
  ANSWER_FIELD_TYPE_ADDRESS = 6;
}

events.proto:

syntax = "proto3";

package com.embroker.oracle;

import "model.proto";

option go_package = "github.com/embroker/oracle/sdk/go_sdk";

// This is the inbound message intended to inform the Oracle of new answers to be persisted
message AnswerUpdateRequest {
  Entity entity = 1;
  repeated Answer answers = 2;
}

// This is the outbound message informing Oracle subscribers of new answers
message AnswersUpdated {
  Entity entity = 1;
  repeated Answer answers = 2;
}