Google Cloud Tasks & Google App Engine Python 3

Question

I am trying to work with the Google Cloud Tasks API

In python2.7 app engine standard you had this amazing library (deferred) that allowed you to easily assign workers to multiple tasks that could be completed asynchronisly.

So in a webapp2 handler I could do this:

create_csv_file(data):
  #do a bunch of work ...
  return

MyHandler(webapp2.Handler):
  def get(self)
    data = myDB.query()
    deferred.defer(create_csv_file, data)

Now I am working on the new Google App Engine Python 3 runtime and the deferred library is not available for GAE Py3.

Is the google cloud tasks the correct solution/replacement?

This is where I am at now... I've scoured the internet looking for answer but my Google powers have failed me. I've found come examples but they are not very good and they appear as though you should be creating creating /adding tasks from gcloud console or locally but no examples of adding tasks from a front end api endpoint.

ExportCSVFileHandler(Resource):
  def get(self):
    create_task()
    return 

CSVTaskHandler(Resource):
  def(post):
    #do a lot of work creating a csv file
    return

create_task():
    client = tasks.CloudTasksClient(credentials='mycreds')
    project = 'my-project_id'
    location = 'us-east4'
    queue_name = 'csv-worker'

    parent = client.location_path(project, location)

    the_queue = {
        'name': client.queue_path(project, location, queue_name),
        'rate_limits': {
            'max_dispatches_per_second': 1
        },
        'app_engine_routing_override': {
            'version': 'v2',
            'service': 'task-module'
        }
    }

    queues = [the_queue]
    task = {
        'app_engine_http_request': {
            'http_method': 'GET',
            'relative_uri': '/create-csv',
            'app_engine_routing': {
                'service': 'worker'
            },
            'body': str(20).encode()
        }
    }

    # Use the client to build and send the task.
    response = client.create_task(parent, task)

    print('Created task {}'.format(response.name))

    # [END taskqueues_using_yaml]
    return response

Answer 1

Yes, Cloud Tasks is the replacement for App Engine Taskqueues. The API can be called from anywhere, ie locally, from App Engine, from external services, and even gcloud. The samples show you how to do this locally, but you can easily replace your old taskqueue code with the new Cloud Tasks library.

Unfortunately, there is no deferred library for Cloud Tasks. There are multiple ways around this. Create separate endpoints for task handlers and use the App Engine routing to send the task to the right endpoint, or add metadata to the task body in order for your handler to appropriate process the task request.