I have a bot which provides information about cinemas like movie schedules and release dates.
Part of this is a cinema_action
trait entity that maps to different values based on the user's input. I have inserted training data for each value.
E.g
cinama_action
--> schedule
cinema_action
--> release_date
However, I am unsure if this is the proposed way of structuring user intents or if I rather should create a single entity for each intent.
The latter would have the advantage of being usable within response and action hints where I can specify which context or entity should be present or not present for an action or response to fire. With the former I can properly group different intents to a single entity, because they all relate to the cinema topic.
With my current setup I can't use cinema_action
as a hint, because the real information lies within the entity's value.