Contextual Rasa

Hi guys, i’m just curious about the contextual rasa. Example if i use DF, there are context input and output, and in rasa, i can’t found anything, and it stores entity in slot. CMIIW, so the rasa story is the same as context in DF, if we define multiple stories regarding 1 flow, so rasa will predict which story to choose from users input. Can someone explain me with this? Thanks!

Rasa NLU does not have any context. You can’t make any context aware intent recognition. At least not without a detour (Providing conversation context to the NLU using microservices). I really hope Rasa will introduce such a feature soon.

Rasa Core helps you to design the “dynamic decision tree”, but it is not an alternative to the Dialogflow Context (Context-aware Dialogue — Rasa NLU 0.12.3 documentation).

1 Like

Hi! I see, i thought it is just the same. Thanks for your explanation! So we have to handle it. So by the meaning of rasa contextual bot is rasa will treat chat history as input to predict the next action, not as a context, but it is similar right? But different.

I’ve just read this docs

I think it is similar with DF context.

Hi Cendekia,

Yes, i’m using that until now to handle everything as per flow, the drawback for that in my opinion is: we can’t change to another intent that has no correlation with current story. Thanks for your response.

1 Like

Agreed about the drawback, that’s why we need to create the story as complete as possible :slight_smile:

Try applying TED policy: Unpacking the TED Policy in Rasa Open Source