Hi , I am confused about the behavior of Rasa. I wrote a story that was not executed correctly. I tried to correct it through interactive learning. But now the response “utter_ask_learning” is called multiple times.
Why is this being done?
My Story:
story: new
steps:
intent: new
action: utter_new
action: utter_ask_thats_help_de -> This action is not always executed . I do not know why
story_ utter_ask_thats_help_de -> if the answeres is no from user
story: story_nicht_geholfen
steps:
intent: deny
action: utter_frage_lernhilfe_de
intent: affirm
action: utter_ask_user
intent: neuer_intent
action: utter_ask_user2
intent: neuer_intent
action: utter_menu_allgemein_de
the follow Story: Wrote with interactive learning:
story: interactive_story_1
steps:
intent: deny
action: utter_ask_for_learning_de Question is displayed 5 times
intent: affirm than follow the nlu_fallback action Why?
action: utter_ask_user
intent: new_intent
action: utter_ask_user2
intent: new_intent
action: utter_menu_allgemein_de
Can someone help me?
I can not see any error
If you want an action to always be executed in response to certain intents, it’s better to make it a rule.
In general, you can think of stories as examples of how your conversations should go. The bot may choose to follow a story even if the conversation so far doesn’t exactly match the story. This allows it to generalize. On the other hand, rules are only followed if the conversation snippet exactly follows the rule, but they are always followed if it does.
I’d recommend you turn your first story (new) into a rule like so:
rules:
- rule: new
steps:
- intent: new
- action: utter_new
- action: utter_ask_thats_help_de
I’m not so sure what’s going wrong with your second story and your interactive story. It looks like it would help to include the action you expect before deny. Could you explain the expected behaviour? Would you like the interactive story to go like the story story_nicht_geholfen?
Hi Felicia
Thanks for your quick answere.
The user asks a question and is offered a solution. Finally, the user is asked whether this solution has helped him. If the answer to the question is no, then the story__nicht_geholfen should start and the user should give a more detailed description and this should then be saved in an intent.
I would set the second story up more like this. You want to give more context than just the intent deny, since your you’re interested only in the deny as a response to utter_ask_thats_help_de