Understanding rasa stories format

In the documentation we have the following example

- story: collect restaurant booking info  # name of the story - just for debugging
  - intent: greet                         # user message with no entities
  - action: utter_ask_howcanhelp
  - intent: inform                        # user message with entities
    - location: "rome"
    - price: "cheap"
  - action: utter_on_it                  # action that the bot should execute
  - action: utter_ask_cuisine
  - intent: inform
    - cuisine: "spanish"
  - action: utter_ask_num_people

My question is simple. Does “rome”, “cheap” and “spanish” - the values for the entities ‘location’ ‘price’ and ‘cuisine’ written in the file stories.yml during training are added to the training examples from the nlu.yml file as new lines in format - <value> at the end of the list of examples:


version: "3.1"
-intent: "inform"
  examples: |
    - [bar](cuisine)
    - [xyz](price)
# added:
    - [rome](location)
    - [cheap](price)
    - [spanish](location)

OR are matched as regexes during predictions and are not used for training?

I can’t find any meaningful explanation for that in the docks.


… added … to the nlu.yml file as new lines


are matched as regexes

Not unless you’ve added them in a lookup table or regex

So why do we write it if they are not doing anything?

stories & rules control the dialog, the nlu.yml is input to the nlu intent and entity recognition. Two separate functions.