New to Rasa, Need Help Understanding

Getting Started with Rasa
1

Hi, I am migrating from Dialogflow to Rasa but I can’t seem to get performance close to what I had from Dialogflow. I believe that a large part of this is due to the context. For my chatbot task, it is designed like a telemarketer, pitching to the user and asking for donations.

In my original Dialogflow model, I relied on many contexts that largely restricted the intent candidates. Additionally, I had two different fallbacks (dependent on the active contexts). For Rasa, the only way for me to replicate these contexts is through stories.md, right? Do I need to list all such possible stories? I have some stories that are very long, is this the right approach? How can I re-create my two different fallbacks?

I am new to Rasa, and I am feeling quite stuck. Any help would be amazing for me. Thank you.

2

@sum1l0st,

In Rasa, the context of a conversation is indeed captured by using stories.

You also capture the context by setting slots, using extracted entities or within custom actions.

If you are new to Rasa, I recommend that you first put together an example bot, as demonstrated in the Masterclass.