Open ArivCR7 opened 4 years ago
Hi @ArivCR7 ,
it is currently not possible to use Rasa for NLU in this context. Alexa will not provide the original user utterances but only the intents as recognized by Alexa NLU.
I'm not sure about this slot mappings. I'm however currently working on a docker based version of this tutorial. I plan to do this within February so please be patient :-)
Hi @BadaBoomi ,
Thanks for the response. I see that slot mappings from Alexa to Rasa are done in EchoNLUMapper class, but I don't see this class being used anywhere. Is this class being used in the code? If so, please point where exactly.
Hi @BadaBoomi @ArivCR7 does anybody able to integrate the Rasa with Alexa?
@svshubham01 , Yes we're able to integrate. Just add this EchoNLUMapper class in the NLU pipeline of config.yml.
Hi @ArivCR7 I want something like whatever i say to alexa comes to my rasa bot and alexa speaks out that rasa response.
I did it NLU and Core part performed at the RASA side @ArivCR7 @BadaBoomi
https://forum.rasa.com/t/google-alexa-connector/60/38?u=svshubham01
Hello @BadaBoomi , Thanks for this awesome tutorial. I wanted to connect my rasa engine with Alexa. I followed the steps here and able to connect!
Here you're using Alexa for NLU and rasa for core. Is it possible to use both core and NLU in rasa and use Alexa only for voice assistant as how we used google assistant in this tutorial: https://blog.rasa.com/going-beyond-hey-google-building-a-rasa-powered-google-assistant/
And the slots in rasa are incorrectly mapped from alexa repsonse. Is EchoNLUMapper class correctly used? I'm not sure where exactly this class is used in the code.
The below are the slots mapped in rasa from alexa:
slot values: {'days': 5693106, 'pillname': '{"type": "IntentRequest", "requestId": "amzn1.echo-api.request.271f6e8a-5a02-4f69-9b8d-16da4d92ef4f", "timestamp": "2020-02-01T13:56:13Z", "locale": "en-US", "intent": {"name": "mood_unhappy", "confirmationStatus": "NONE"}}', 'requested_slot': 'schedule', 'schedule': '{"type": "IntentRequest", "requestId": "amzn1.echo-api.request.39272921-497a-4d4e-90ae-99a5b0a0df50", "timestamp": "2020-02-01T13:56:46Z", "locale": "en-US", "intent": {"name": "mood_unhappy", "confirmationStatus": "NONE"}}', 'time': ['2020-02-01T00:00:00.000+05:30', '2020-02-02T13:56:46.000+05:30']}