RegexEntityExtractor fills the slot only with the latest match

Rasa Open Source
1

Suppose I have an intent like:

- intent: is_employee_busy
    examples: |
      - [John Smith](name) is busy?
      - [John Smith](name) is what doing?

- regex: name
  examples: |
    - [A-Za-z]+\s[A-Za-z]+

With the default pipeline turned on regex and case insensitive. Btw, don’t think about the grammar of the intents, the actual examples are in another language, so changing the order of the words is not the case. So, the problem is that extractor parses the intent like [John Smith](name) [is busy](name), so that the slot name is filled with is busy. How do I take only the first match or save all the matches?

1 Like
2

bump. Any help? Had to turn on case sensitive for now

3

For english, you should try Spacy as discussed here.