Entity being extracted by multiple entity extractors breaks testing

Rasa Open Source
1

I am using an regex entity extractor explained here RASA Regex Entity Extraction - Naoko - Medium

Sometimes DIET extracts the same entity as the regexEntityExtractor. In the form action this is not an issue because I try to map the regexEntity before the DIET entity to a slot.

But for testing this is a problem because I can not specify two entity names for one word. This is where it fails

* inform: [bouwjaar](property) <!-- predicted: inform: [bouwjaar](property_regex_extractor)[bouwjaar](property) -->

Is there a solution for this problem?

2

Hey @superrroy

Sorry, I didn’t understand what you’re trying to archive, could you please define the goal?

3

Already found the problem. I was also training the entity in the nlu which shouldn’t happen because I was already extracting it using the custom entity extractor

4

@superrroy How did you solve this peoblem