Here is an example of an End to End test that I put in my tests/conversation_tests.md file.
## Foreign language story
* greet: hi
- action_greet
- rewind
* ask_book: Are there any books in [french](foreign_language)?
- utter_ask_book
When I run rasa test
, I get the following result in results/failed_stores.md:
## Foreign language story
* greet: hi
- action_greet
- rewind
* ask_book: Are there any books in [french](foreign_language)? <!-- predicted: ask_book: Are there any books in [french](foreign_language)[french](foreign_language) -->
- utter_ask_book
I think what’s happening is that I have two entity extractors (DIETClassifier and CRFEntityExtractor) that are both independently finding the ‘french’ foreign_language entity, so it gets duplicated when the test is run. I’m worried this double entity extraction could be affecting my model training. Is there a way to change this behavior?
Thanks