Apply Sentiment Analysis through a Slot

Rasa Open Source
1

Hi , i need to apply sentiment analysis to a slot called feedback from getting user provided text through the form and the bot will need to utter different type of message based on the sentiment values (pos, neu , neg) through custom action e.g through the dispatcher.utter_message(template=“utter_based_on_sentiment”)

can anyone provide me with reference or guide me how can i do it ? I found a reference from post but i not sure how to implment which post by : JulianGerhard

Thanks

2

Can anyone help me on this ? I had already apply Sentiment component to my chatbot , currently my implementation was

slots:
  sentiment:
    type: categorical
    values:
      - positive
      - neutral
      - negative

responses:
  utter_based_on_sentiment:
    - condition:
        - type: slot
          name: sentiment
          value: positive
      text: "Hey, {name}. Nice to see you again! How are you?"

    - condition:
        - type: slot
          name: sentiment
          value: neutral
      text: "Hey, {name}. How are you?"

Above using conditional response variations .

actions.py

Class ActionSubmitFeedback(Action):

    def name(self) -> Text:

        return "action_submit_feedback"

  

    def run(

        self,

        dispatcher: CollectingDispatcher,

        tracker: Tracker,

        domain: Dict[Text, Any],

    ) -> List[Dict]:

        sentiment = tracker.get_slot("sentiment")

        dispatcher.utter_message(template="utter_based_on_sentiment")

i have a form in rules.yml which flow to ActionSubmitFeedback Action Class when form is filled with a feedback slot to save as slot from user text . How can i change my code in action to let my chatbot utterance different messge based on the feedback slot .

3

Have you seen our new blogpost on conditional response variation?

4

Hi koaning yes , i did and also implemented but didn’t show here , i think i have i solved my problem which i asked. Current now i faced another challenge is to obtain the slot value the sentiment that was predicted by my custom component and i not sure is this the right way to do it and i tested by running the action to pass to utterance but it cache my old response >.< i does not know why

actions.py
sentiment = tracker.get_slot("sentiment")

dispatcher.utter_message(template="utter_display_sentiment", customer_sentiment = sentiment) 

domain.yml
utter_display_sentiment:

  - text: Thanks {customer_sentiment}
5

but it cache my old response

Could you clarify what you mean with this statement? What is happening in the conversation and what did you expect instead?

6

can you please share your nlu file?