FormAction slot was requested but nothing was extracted it will allow other policies to predict another action.
But I want this, It should not allow to predict untill slots does not get filled.
you need to override validate and do not raise Exception
Hi @Ghostvv I need your help please I want to use forms for a specific case I have and I can’t get it to work.
In actions as a form I have :
class IdeasmineForm(FormAction):
"""Accept free text input from the user for suggestions"""
def name(self):
return "ideasmine_form"
@staticmethod
def required_slots(tracker):
return ["ideasmine_message"]
def slot_mappings(self):
return {"ideasmine_message": self.from_text()}
def submit(self, dispatcher, tracker, domain):
return[]
So I want to take the input of the user as it is, I know the user will give me the intent : ideasmine_message but I don’t know what it will be it’s very random so in my intents I can’t give examples for this intent. This is why I want to use forms to just get the user input and take this as ideasmine_message intent/slot
After the FormAction I want to tell the user simply that he taped {ideasmine_message[text]}, so in my stories I have the next action like:
class action_SendIntent_B_S(Action):
def name(self):
return 'action_SendIntent_B_S'
def run(self, dispatcher, tracker, domain):
loc = tracker.get_slot('ideasmine_message')
response = """Hi user you said {} , """.format(loc['text'])
dispatcher.utter_message(response)
Just tell the user what he taped.
stories.md
happy path
-
bonjour
- utter_bonjour
-
confirmer
- ideasmine_form
- form{“name”: “ideasmine_form”}
- form{“name”: null}
- action_SendIntent_B_S
in domain.yml
- utter_ask_ideasmine_message
- action_etape4
forms:
- ideasmine_form
intents:
- confirmer
- ideasmine_message
entities:
- ideasmine_message
slots:
ideasmine_message:
type: unfeaturized
auto_fill: false
templates:
utter_ask_ideasmine_message:
- text: What is your Idea?
utter_bonjour:
- text: hi, do you need help?
What I realy want is:
user: Hi (*bonjour)
bot: hi, do you need help? (-utter_bonjour)
user: yes (*confirmer)
bot: what is your idea? (I could of use this in stories as *ideas_message but the problem is that I don't know what should I receive as an answer it could be any thing, this why I want to use forms. So the bot will execute ideasmine_form which will show the utter_ask_ideasmine_message)
user: my idea is my idea (filling the form and taking this sentence as ideasmine_message slot)
bot: you said: my idea is my idea (action_SendIntent_B_S)
when I tried this with interactive learning the bot predict well till the ideamine_form (correct?) I tape Yes, then the bot show the utter_ask_ideamine_message and propose to listen I tape Yes and taped the idea (my idea is my idea) Here it stops because the bot propose to classify this as None intent and if I tell him to classify it as ideamine_message he re do the ideamine_form as the new action.
Am I doing something wrong ? I’am using the last versions of both rasa_nlu and rasa_core
I didn’t understand, what is the problem?
How would I do that ? Do I have to changed the conditions ?? I did not get you.
The embadding policy just jumps over the Form and keras policy is a liltle good but also jumps when it could not fill the Form.
remove this: rasa_core_sdk/forms.py at 7f91132aecda9c4ea6f1c97a58a52ed5a7178efe · RasaHQ/rasa_core_sdk · GitHub
1 Like
These line of code ??
if not slot_values:
# reject to execute the form action
# if some slot was requested but nothing was extracted
# it will allow other policies to predict another action
raise ActionExecutionRejection(self.name(),
"Failed to validate slot {0} "
"with action {1}"
"".format(slot_to_fill,
self.name()))
yes