Entities can't get extracted with regex

Rasa Open Source
1

Hey guys, how are you? I need help, probaly a missed something about entities extractor with regex, i’ve try everthing that i found on the docs but nothing work, sorry about my english, i’m no native speaker but i will take this opportunity to risk write a litle bit hahah pls, look this. My stories:

1

My nlu:

2

My domain:

3
3313×715 28.4 KB

This runs perfect, or almost. I need catch this numbers callled “cpf” and put them on the cpf_slot, the intention predict its look like a divine job, its super higth, but the entity don’t come together, look this photo:

5
5804×520 40.8 KB

more than 99% of similarity in the intention but nothing about the entities? I know that, the classification and the extraction methods its a diferent thing, but this sounds very strange to me. Anyone can help me please? I don’t have hair anymore :cry:

2

The “rule” wrote in the first photo is not the problem, i’ve try everthing, with rules and stories hahah, plis, ignore that…

3

Hello and Welcome to the forum @Wedrano-de-Carvalho.

can you please share the config.yml file and rasa --version?

4

Yaaaahhh, hellooo o/

shure, my config file its all commented, i’m using the RASA default configs…

6
61116×876 100 KB

and the version

7

5

one more observation, in the first photo the entity name was diferent, this also isn’t the problem, i already changed that

6

Hey, i completed the discobot tutorial and i see that i have to mark the peaples with the ‘@’ hahah so i try it marking you @nik202 auhsaushuashu

7

@Wedrano-de-Carvalho can you uncomment the config.yml file and try run?

8

@nik202 Didn’t work :smiling_face_with_tear:

9

I trained again and nothing change

10

@Wedrano-de-Carvalho delete older trained model please and try again. Did you mentioned the RegexEntityExtractor in config.yml please ref this link: NLU Training Data

11

@nik202 I already have delete the other models, now i’m replace RegexFeaturizer to RegexEntityExtractor, but when i run rasa train command, this messages is printed:

11
11727×106 22.8 KB
but when i run the regex in jupyter notebooks, it’s all right: 10

12

@Wedrano-de-Carvalho you trained the model successfully?

13

@nik202 yes, but that message is printed every time i run the rasa, and the entity is not extracted

14

@Wedrano-de-Carvalho do you have nlu examples based on your regex?

15

@nik202 what do you mean with that? in my nlu.py file? if that’s what you mean, i have: 2

16

@nik202 hey, i got it!! yohohohohoh i changed the nlu text format to this 12

now its work!! thank’s a lot for you attention bro, you are the best!!

17

@Wedrano-de-Carvalho please close this thread please as a solution while Entities can't get extracted with regex - #14 by nik202 , it will be 250 :slight_smile: and good luck!

18

@nik202 I don’t know how i close this topic kk

19

@Wedrano-de-Carvalho Thanks you marked the right solution.