Rasa: Rasa X - annotate new data - doubles

[PhilippeBaert]

Rasa version: 1.0.1

Rasa X version: 0.19.0

Python version: 3.6.7

Operating system (windows, osx, ...): Windows 10

Issue:

Hello,

Imagine the following flow:
1) I talk to the bot through Interactive learning and save a couple of stories
2) I go to Training => NLU Training => Annotate new data
3) I want to add new sentences I generated in interactive learning to my NLU training data

Issue: the sentences I used in interactive learning, are added twice.
Once for the NLU prediction, once for the intent that I set through interactive learning.

You will see for example, the first sentence: "Am I talking to a chatbot?"
This was wrongly classified as "goodbye". So I manually changed it to "ask_isbot" in interactive learning.

In my opinion, if both the NLU predicted intent, and the intent I set through interactive learning are the same, it shouldn't be added twice to the annotate screen.

Thanks for looking into this.

Cheers,
Philippe

[akelad]

Thanks for raising this issue -- we'll look into fixing it for the next release

[tmbo]

I think if you correct something in interactive learning, it should be added to the training data and shouldn't be shown in the inbox at all.

[wochinge]

@TyDunn Why is this reopened?

[TyDunn]

My apologies. I was adding it to a ZenHub milestone board and did not realize moving it to the done column would reopen it. Just closed it. Sorry for the noise @wochinge

[wochinge]

No worries 😄