Fix test_stories.md (Rasa 1.10.x)

This question is pertained to Rasa 1.10.x. In my test_stories.md, I have an entry like this:

* Execute_job: Hi team, How to execute a spark job?
  - utter_execute_job_spark

After running rasa test -s test_stories.md, I got this error in failed_stories.md:

* Execute_job: Hi team, How to execute a spark job?   <!-- predicted: Execute_job: Hi team, How to execute a [spark](platform) job? -->
    - slot{"platform": "spark"}
    - utter_execute_job_spark

My question is how to correct the entry in stories.md so that this entry will pass the test. I tried below but still failed.

* Execute_job{"platform": "spark"}: Hi team, How to execute a spark job?
  - utter_execute_job_spark

It has taken me a while to find a possible solution, because I am not used to see or use the 1.x version of Rasa.

I think the problem may be that you must specify the entity that is being extracted from the text in the text itself:

* Execute_job: Hi team, How to execute a [spark](platform) job?
  - utter_execute_job_spark

You can see this structure in the following link Testing Your Assistant, in the Basics section and in the Forms Happy Path section:

image

image

In this way the stories are correctly written in the end-to-end format, including both the NLU output and the original text.

In the second option you propose, you should only specify the slot that way if it has been added from a custom action:

image

I hope the solution helps, if not, I hope someone else can lend you a hand. :wink: