Getting warning in rasa_core.processor

Hey guys,

I am getting a warning in logs when different actions are executed. Can you please explain, should I ignore this or need to dealt with.

2018-11-22 18:54:45 DEBUG    rasa_core.processor  - Received user message 'i live in banglore' with intent '{'confidence': 0.9592106342315674, 'name': 'store_search'}' and entities '[{'confidence': 0.9788179034608674, 'extractor': 'ner_crf', 'start': 10, 'entity': 'location', 'value': 'banglore', 'end': 18}]'
2018-11-22 18:54:45 DEBUG    rasa_core.processor  - Logged UserUtterance - tracker now has 267 events
2018-11-22 18:54:45 DEBUG    rasa_core.processor  - Current slot values: 
        product: None
        location: banglore
        store_id: None
        zone_id: None
        price: None
2018-11-22 18:54:45 DEBUG    rasa_core.policies.memoization  - Current tracker state [{'slot_location_0': 1.0, 'prev_utter_general_order_enquiry': 1.0, 'intent_general_order_enquiry': 1.0}, {'slot_location_0': 1.0, 'entity_location': 1.0, 'intent_order': 1.0, 'prev_action_listen': 1.0}, {'slot_location_0': 1.0, 'prev_action_store_location': 1.0, 'entity_location': 1.0, 'intent_order': 1.0}, {'slot_location_0': 1.0, 'prev_action_search_location': 1.0, 'entity_location': 1.0, 'intent_order': 1.0}, {'slot_location_0': 1.0, 'intent_store_search': 1.0, 'entity_location': 1.0, 'prev_action_listen': 1.0}]
2018-11-22 18:54:45 DEBUG    rasa_core.policies.memoization  - There is no memorised next action
2018-11-22 18:54:45 DEBUG    rasa_core.policies.ensemble  - Predicted next action using policy_1_KerasPolicy
2018-11-22 18:54:45 DEBUG    rasa_core.processor  - Predicted next action 'action_store_location' with prob 1.00.
2018-11-22 18:54:45 DEBUG    rasa_core.actions.action  - Calling action endpoint to run action 'action_store_location'.
2018-11-22 18:54:45 DEBUG    rasa_core.processor  - Action 'action_store_location' ended with events '['SlotSet(key: location, value: banglore)', 'SlotSet(key: zone_id, value: None)', 'SlotSet(key: product, value: None)', 'SlotSet(key: store_id, value: None)']'
2018-11-22 18:54:45 WARNING  rasa_core.processor  - Action 'action_store_location' set a slot type 'zone_id' that it never set during the training. This can throw of the prediction. Make sure to include training examples in your stories for the different types of slots this action can return. Remember: you need to set the slots manually in the stories by adding '- slot{"zone_id": null}' after the action.
2018-11-22 18:54:45 WARNING  rasa_core.processor  - Action 'action_store_location' set a slot type 'product' that it never set during the training. This can throw of the prediction. Make sure to include training examples in your stories for the different types of slots this action can return. Remember: you need to set the slots manually in the stories by adding '- slot{"product": null}' after the action.
2018-11-22 18:54:45 WARNING  rasa_core.processor  - Action 'action_store_location' set a slot type 'store_id' that it never set during the training. This can throw of the prediction. Make sure to include training examples in your stories for the different types of slots this action can return. Remember: you need to set the slots manually in the stories by adding '- slot{"store_id": null}' after the action.

Thanks

this warning means that your custom actions return SetSlot events that you didn’t specify in the training stories, if these slots are unfeaturized you’re fine, otherwise please use interactive training to see how the training story for this custom action should look like

Im getting the same error and I used interactive learning to create a training story for this, retrained the model, and the warning is still coming up.

do your action set a slot that is not present in the stories as SlotSet event after this action?

I am getting the same error. Please clarify. @akelad @Juste @Ghostvv

what slots your custom action returns? how the story for this action looks like?

Hey @rohitharitash I actually found the same problem happening in my code and resolved it by setting slot in the code.

`

  • action: restaurant_form
    • active_loop: restaurant_form
    • slot_was_set:
      • cuisines: Indian
      • times: 8:00pm`

How about when you’re using a rule and not a story?

At first I had both a rule and a story, and got the “conflicting stories” error, so I removed my story to just use the rule, and now it errors wanting the slot to be filled during training.

I see no way to do that in a rule because as a rule you probably don’t want to force a slot to set every time it is executed, right?

I have the same issue here. I use rules to handle the form and not stories. And I get the error.

So what is the proper way to handle this? Form by rules or stories?