How to handle incoming PubSub messages in Python?

I have created a Cloud Compute Engine instance on Debian, and have successfully created a PUSH subscription to a topic with

from google.cloud import pubsub_v1

project_id = "censored"
topic_name = "censored"
subscription_name = "censored"
endpoint = "https://censored.appspot.com/pubsub/push?token=censored"

def create_push_subscription(project_id,
                             topic_name,
                             subscription_name,
                             endpoint):
    """Create a new push subscription on the given topic."""
    # [START pubsub_create_push_subscription]

    subscriber = pubsub_v1.SubscriberClient()
    topic_path = subscriber.topic_path(project_id, topic_name)
    subscription_path = subscriber.subscription_path(
        project_id, subscription_name)

    push_config = pubsub_v1.types.PushConfig(
        push_endpoint=endpoint)

    subscription = subscriber.create_subscription(
        subscription_path, topic_path, push_config)

    print('Push subscription created: {}'.format(subscription))
    print('Endpoint for subscription is: {}'.format(endpoint))
    # [END pubsub_create_push_subscription]

create_push_subscription(project_id, topic_name, subscription_name, endpoint)

but I'm not sure how exactly incoming messages arrive. I have found this sample code to parse messages, but I'm not sure how to get it to monitor in the background and 'activate' whenever incoming messages arrive.

import argparse
import base64
import json
import sys
import time

from google.cloud import pubsub_v1

def summarize(message):
    # [START parse_message]
    data = message.data.decode('utf-8')
    attributes = message.attributes

    name = attributes['name']
    time_created = attributes['timeCreated']
    bucket_id = attributes['bucketId']
    object_id = attributes['objectId']
    generation = attributes['objectGeneration']
    description = (
        '\tName: {name}\n'
        '\tTime Created: {time_created}\n'
        '\tBucket ID: {bucket_id}\n'
        '\tObject ID: {object_id}\n'
        '\tGeneration: {generation}\n'
        ).format(
            name=name,
            time_created=time_created,
            bucket_id=bucket_id,
            object_id=object_id,
            generation=generation
            )

    if 'overwroteGeneration' in attributes:
        description += '\tOverwrote generation: %s\n' % (
            attributes['overwroteGeneration'])
    if 'overwrittenByGeneration' in attributes:
        description += '\tOverwritten by generation: %s\n' % (
            attributes['overwrittenByGeneration'])

    payload_format = attributes['payloadFormat']
    if payload_format == 'JSON_API_V1':
        object_metadata = json.loads(data)
        name = object_metadata['name']
        time_created = object_metadata['timeCreated']
        size = object_metadata['size']
        content_type = object_metadata['contentType']
        metageneration = object_metadata['metageneration']
        description += (
            '\tName: {name}\n'
            '\tTime Created: {time_created}\n'
            '\tContent type: {content_type}\n'
            '\tSize: {object_size}\n'
            '\tMetageneration: {metageneration}\n'
            ).format(
                name=name,
                time_created=time_created,
                content_type=content_type,
                object_size=size,
                metageneration=metageneration
                )
    return description
    print('Note for developer: If BucketId and ObjectId listed, utf encoding.')
    print('If not, JSON_V1 encoding. Adjust accordingly.')

    # [END parse_message]
while(True):
    print("signpost 1")
    summarize(message)
    print("signpost 2")
    time.sleep(10)
print("signpost 3")

For example, this code will return

NameError: name 'message' is not defined

which is expected...

Could someone please help me set it up properly?

I know it's different in PULL because then the message will be defined during the pull, but I'd like to keep it as PUSH, if possible.