when u do form actions the bot start asking question from domain file where there is “utter_ask_slot name”…but i want the bot to ask from actions where “action_ask_slot name” some thing like this
When u do form actions the bot start asking question from domain file where there is "utter_ask_slot name"....but i want the bot to ask from actions where "action_ask_slot name" some thing like this
Hi @faiza_conte. I am not sure I follow. Could you try explaining it again with an example?
To do formaction or slot filling the bot askes the question from utter_ask_slotname but I want the bot to ask from action_ask_slotname
@faiza_conte Is there a reason you want to call a custom action to send the message as opposed to letting the FormAction send the utterance?
Yes…for example in a case of where there is a button selection it could be from utter but is the buttons are fetched from database then there has to be some action to be done from the backend so in that case I have to do actions
Hi @faiza_conte, you can trigger your custom action of choice in the form by overwriting the
request_next_slot method in your Form action. We have an example of how to do that in Rasa 1.10 here under
okay thanks…and another question do u know about data cleaning???and how rasa does data cleaning??
@faiza_conte Rasa will perform some data validation checks when running
rasa train, but won’t perform any data cleaning for you. You’ll have to review your NLU & story data before training to make sure the data is clean.
so it is me that manually have to go through my data to clean and identify if there are any errors found??
@faiza_conte You can run
rasa data validate (Validate Data) to identify errors (
rasa train will automatically run this as well), but you will need to clean the data yourself.
so specifically i will have to do it anyways,but what about the piplines …are they for data validation or other purposes??
@faiza_conte The pipeline is for performing NLU intent classification and entity extraction. You can read more about creating an NLU pipeline here.
so none of them is related to data cleaning?