Rasa.core.agent import Agent Issue

Rasa Open Source
1

I am using Rasa Version 2.1.0 and I seem to be getting an issue when I import Agent. This is the error I receive:

TypeError Traceback (most recent call last)

[<ipython-input-70-8f67cf765ed0>](https://localhost:8080/#) in <module>() ----> 1 from rasa.core.agent import agent 2 3 # loading the Core training samples 4 ##training_data = load_data("stories.md") 5

9 frames

[/usr/local/lib/python3.6/dist-packages/rasa/core/agent.py](https://localhost:8080/#) in <module>() 27 from rasa.core.lock_store import InMemoryLockStore, LockStore 28 from rasa.core.nlg import NaturalLanguageGenerator ---> 29 from rasa.core.policies.ensemble import PolicyEnsemble, SimplePolicyEnsemble 30 from rasa.core.policies.memoization import MemoizationPolicy 31 from rasa.core.policies.policy import Policy, PolicyPrediction

[/usr/local/lib/python3.6/dist-packages/rasa/core/policies/__init__.py](https://localhost:8080/#) in <module>() 3 4 # and after that any implementation ----> 5 from rasa.core.policies.ensemble import SimplePolicyEnsemble, PolicyEnsemble

[/usr/local/lib/python3.6/dist-packages/rasa/core/policies/ensemble.py](https://localhost:8080/#) in <module>() 38 from rasa.core.policies.fallback import FallbackPolicy 39 from rasa.core.policies.memoization import MemoizationPolicy, AugmentedMemoizationPolicy ---> 40 from rasa.core.policies.rule_policy import RulePolicy 41 from rasa.shared.core.trackers import DialogueStateTracker 42 from rasa.shared.core.generator import TrackerWithCachedStates

[/usr/local/lib/python3.6/dist-packages/rasa/core/policies/rule_policy.py](https://localhost:8080/#) in <module>() 41 from rasa.shared.core.domain import InvalidDomain, State, Domain 42 from rasa.shared.nlu.constants import ACTION_NAME, INTENT_NAME_KEY ---> 43 import rasa.core.test 44 import rasa.core.training.training 45

[/usr/local/lib/python3.6/dist-packages/rasa/core/test.py](https://localhost:8080/#) in <module>() 10 from rasa.shared.exceptions import RasaException 11 import rasa.shared.utils.io ---> 12 from rasa.core.channels import UserMessage 13 from rasa.shared.nlu.training_data.message import Message 14 from rasa.shared.core.training_data.story_writer.yaml_story_writer import (

[/usr/local/lib/python3.6/dist-packages/rasa/core/channels/__init__.py](https://localhost:8080/#) in <module>() 16 from rasa.core.channels.botframework import BotFrameworkInput # nopep8 17 from rasa.core.channels.callback import CallbackInput # nopep8 ---> 18 from rasa.core.channels.console import CmdlineInput # nopep8 19 from rasa.core.channels.facebook import FacebookInput # nopep8 20 from rasa.core.channels.mattermost import MattermostInput # nopep8

[/usr/local/lib/python3.6/dist-packages/rasa/core/channels/console.py](https://localhost:8080/#) in <module>() 6 7 import aiohttp ----> 8 import questionary 9 from aiohttp import ClientTimeout 10 from prompt_toolkit.styles import Style

[/usr/local/lib/python3.6/dist-packages/questionary/__init__.py](https://localhost:8080/#) in <module>() 3 4 import questionary.version ----> 5 from questionary.form import Form 6 from questionary.form import form 7 from questionary.prompt import prompt

[/usr/local/lib/python3.6/dist-packages/questionary/form.py](https://localhost:8080/#) in <module>() 1 from collections import namedtuple 2 ----> 3 from questionary.constants import DEFAULT_KBI_MESSAGE 4 from questionary.question import Question 5

[/usr/local/lib/python3.6/dist-packages/questionary/constants.py](https://localhost:8080/#) in <module>() 39 ("selected", ""), # style for a selected item of a checkbox 40 ("separator", ""), # separator in lists ---> 41 ("instruction", ""), # user instructions for select, rawselect, checkbox 42 ] 43 )

TypeError: object() takes no parameters

Please help me out I am stuck. I’ve tried changing the Rasa version and still nothing.

2

Hi @IsraaMohamedHamid. When do you get this error? Is it when you run rasa train? Do you get the same issue when you are just initialising new Rasa project with rasa init?

3

I’m using google Colab. It appears after this code:

from rasa.core.interpreter import RasaNLUInterpreter 
from rasa.core.policies.policy import *
from rasa.core.agent import Agent
from rasa.core import training

agent = Agent('domain.yml', policies=[MemoizationPolicy(),  MappingPolicy(), TEDPolicy()])

# Configuration for Rasa Core.

# https://rasa.com/docs/rasa/core/policies/

# loading our neatly defined training dialogues

training_data = agent.load_data('stories.md')

agent.train(

    training_data,

    validation_split=0.2,

    epochs=200, 

    max_history= 5
)

agent.persist('models/dialogue')
4

Hi @IsraaMohamedHamid. It appears to me that you are using a snipped of code written for an old version of Rasa. Could you test your assistant in a command line and use rasa train function for training the model and rasa shell for testing an assistant?

5

I’ve tried in using cmd line and it works however I need it work for colab as well so I was wondering what the alternative of Agent is for the new version of Rasa? Also I’ve tired using this code rasa.core.train('domain.yml', 'stories.md', 'rasa_Core_models') and I’m not sure I think it trainings but it doesn’t save the result.

6

@IsraaMohamedHamid Hi, have you solved this issue? I am facing the same issue too, and I need it to work on Colab too TT