Rasa X removing regex from nlu.md

[Deprecated] Rasa X Community Edition
1

Rasa version: 1.10.0 Rasa X version: 0.28.5

While making some trainning on personal project, I’ve seem that rasa x consistently removes my regex mention on my trainning data. My nlu data looks like that on its end:

...

## lookup:client_first_name.csv
  data/names/client_first_name.csv

## lookup:city.txt
  data/nlu/lookups/city.txt

## lookup:course.txt
  data/nlu/lookups/course.txt

## regex:cpf
- ([0-9]{2}[\.]?[0-9]{3}[\.]?[0-9]{3}[\/]?[0-9]{4}[-]?[0-9]{2})|([0-9]{3}[\.]?[0-9]{3}[\.]?[0-9]{3}[-]?[0-9]{2})

Am I missing something here?

2

Hello @lgabs ,

your stories does make use of this regexp? Perhaps it’s removing because isn’t begin used.

Cya!

3

@lluchini, thanks for the reply. I’m using them on stories, such as this simple example:

## the buyer
* already_bought
    - utter_ask_info_to_buyer
* inform_cpf{"cpf":"58769397147"}
    - cpf_form
    - form{"name":"cpf_form"}
    - slot{"cpf":"58769397147"}
    - form{"name":null}
    - slot{"requested_slot":null}
    - utter_forwarding
    - action_human_silent_handoff

There is also a cpf entity defined on domain.yml. However, I’m afraid this is not being used sice I’m not using CRFEntityExtractor on the pipeline, I thought DIETClassifier would do it.