rasa: TypeError: 'coroutine' object is not iterable

Rasa version: rasa 1.0.2
rasa-core-sdk 0.14.0
rasa-sdk 1.0.0

Python version: Python 3.7.3

Operating system (windows, osx, …): Mac osx Mojave

Issue: When I run the below command: python dialogue_management_model.py

The error that pops up is this frame below:

WARNING: The TensorFlow contrib module will not be included in TensorFlow 2.0. For more information, please see:

https://github.com/tensorflow/community/blob/master/rfcs/20180907-contrib-sunset.md
https://github.com/tensorflow/addons If you depend on functionality not listed there, please file an issue.

Traceback (most recent call last): File “dialogue_management_model.py”, line 38, in <module> train_dialogue() File “dialogue_management_model.py”, line 25, in train_dialogue agent.train(data) File “/usr/local/lib/python3.7/site-packages/rasa/core/agent.py”, line 668, in train self.policy_ensemble.train(training_trackers, self.domain, **kwargs) File “/usr/local/lib/python3.7/site-packages/rasa/core/policies/ensemble.py”, line 89, in train policy.train(training_trackers, domain, **kwargs) File “/usr/local/lib/python3.7/site-packages/rasa/core/policies/memoization.py”, line 152, in train for t in training_trackers TypeError: ‘coroutine’ object is not iterable sys:1: RuntimeWarning: coroutine ‘Agent.load_data’ was never awaited

I’m unsure of how this is happening, and I saw on another post to add “await” before when i train the agent, but that gave a syntax error. Please help!!

Content of configuration file (config.yml):

{
    "pipeline":"spacy_sklearn",
    "path":"./models/nlu",
    "data":"./data/data.json"
}

Content of domain file (domain.yml):

slots:
  location:
    type: text


intents:
  - greet
  - goodbye
  - inform


entities:
  - location

templates:
  utter_greet:
    - 'Hello! How can I help you today?'
    - 'Hi there!'
  utter_goodbye:
    - 'Goodbye! I hope I was of assistance today :)'
    - 'Bye bye!'
    - 'See you later!'
  utter_ask_location:
    - 'Could you specify the location?'
    - 'In what location?'
    - 'Where, exactly?'


actions:
  - utter_greet
  - utter_goodbye
  - utter_ask_location
  - action_weather

Content of dialogue_management_model.py:

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
from __future__ import unicode_literals

import logging
import rasa.core
from rasa.core.agent import Agent
from rasa.core.policies.keras_policy import KerasPolicy
from rasa.core.policies.memoization import MemoizationPolicy
from rasa.core.interpreter import RasaNLUInterpreter
from rasa.core.run import serve_application
from rasa.core import config

logger = logging.getLogger(__name__)

def train_dialogue(domain_file = 'weather_domain.yml',
					model_path = './models/dialogue',
					training_data_file = './data/stories.md'):
					
	agent = Agent(domain_file, policies = [MemoizationPolicy(), KerasPolicy(max_history=3, epochs=200, batch_size=50)])
	data = agent.load_data(training_data_file)	
	

	agent.train(data)
	
	agent.persist(model_path)
	return agent
	
def run_weather_bot(serve_forever=True):
	interpreter = RasaNLUInterpreter('./models/nlu/default/weathernlu')
	agent = Agent.load('./models/dialogue', interpreter=interpreter)
	rasa.core.run.serve_application(agent ,channel='cmdline')
		
	return agent
	
if __name__ == '__main__':
	train_dialogue()
	run_weather_bot()

About this issue

Original URL
State: closed
Created 5 years ago
Reactions: 3
Comments: 16 (7 by maintainers)

Most upvoted comments

We use asyncio internally. agent.load_data is a coroutine that needs to be run and waited for. Please try out the following code:

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
from __future__ import unicode_literals

import logging
import asyncio
import rasa.core
from rasa.core.agent import Agent
from rasa.core.policies.keras_policy import KerasPolicy
from rasa.core.policies.memoization import MemoizationPolicy
from rasa.core.interpreter import RasaNLUInterpreter
from rasa.core.run import serve_application
from rasa.core import config

logger = logging.getLogger(__name__)


async def train_dialogue(domain_file='weather_domain.yml',
                   model_path='./models/dialogue',
                   training_data_file='./data/stories.md'):
    agent = Agent(domain_file, policies=[MemoizationPolicy(),
                                         KerasPolicy(max_history=3, epochs=200,
                                                     batch_size=50)])

    loop = asyncio.get_event_loop()
    data = loop.run_until_complete(agent.load_data(training_data_file))

    agent.train(data)

    agent.persist(model_path)
    return agent


def run_weather_bot(serve_forever=True):
    interpreter = RasaNLUInterpreter('./models/nlu/default/weathernlu')
    agent = Agent.load('./models/dialogue', interpreter=interpreter)
    rasa.core.run.serve_application(agent, channel='cmdline')

    return agent


if __name__ == '__main__':
    train_dialogue()
    run_weather_bot()

As you are using python3.7 you should be able to replace

    loop = asyncio.get_event_loop()
    data = loop.run_until_complete(agent.load_data(training_data_file))

    asyncio.run(agent.load_data(training_data_file))

tabergma on Jun 4, 2019

@neerajb1 As soon as you put the keyword async to the function definition, it becomes a coroutine. If you not await the coroutine, the error you mentioned will show up.

In your case you can simply remove the keyword async from train_dialogue as it is not needed. Please have a look at https://docs.python.org/3/library/asyncio-task.html#coroutine for more information.

Try the following:

def train_dialogue(
  domain_file = 'domain.yml',
  model_path = 'models/dialogue',
  training_data_file = 'data/stories.md'
):

    agent = Agent(domain_file, policies = [MemoizationPolicy(), KerasPolicy(max_history=3, epochs=200, batch_size=50)])

    loop = asyncio.get_event_loop()
    data = loop.run_until_complete( agent.load_data(training_data_file ))

    agent.train(data)
    agent.persist(model_path)
    return agent

tabergma on Jun 14, 2019

@tabergma, thank you! that solved the problem!

naoko on Jun 4, 2019