Hey,
I have a slot of type list set in my domain.
slots:
key_features:
type: list
When i extract the value of the slot in my code
tracker.get_slot(“key_features”)
I get a list of the features which is the expected behavior, now when i use the FormAction implementation things get unexpected as i walk through the interactive learning process
class ActionDummyClass(FormAction):
RANDOMIZE = False
@staticmethod
def required_fields():
return [
EntityFormField("key_feature", "key_features"),
....
]
def name(self):
return 'action_dummy_class'
after each input, the nlu is shown to correctly detect the entities and put them in a list as expcted
slot{“key_features”: [“feature 1”, “feature 2”]}
but when the action class is called
action_dummy_class
only the first value of the list is recorded and shown as type text
slot{“key_features”: “feature 1”}
I went through the code to see how EntityFormField is implemented.
in the extract method
value = next(tracker.get_latest_entity_values(self.entity_name), None)
is there a reason for calling next only once and just taking the first value?