'ended' should auto fire when there is no voice response about google-assistant

[endoplasmic]

Solid find!

I went deep digging, but ended up just triggering the end() method whenever we need to change the volume.

[ppisljar]

I am sure there are more cases like this, so detecting google's response would be the right way to go.
imagine something like google, turn on the light (light comes on, no voice response ....)

[endoplasmic]

Does that happen for you? Whenever I do it, she always replies with something like "Okay, turning on xxxxx light" or "Turning on 2 lights", something like that.

Yah, I agree with you though, I need to find out if the python version handles it differently, but I have to admit that my python skills are pretty rough.

[ppisljar]

one (not the best) way would be with a timeout (if in 100ms after end of utterance first audio packet is not received we end the conversation) ....

[endoplasmic]

If you wanted to implement something like that, you could do it on your server.

Start a timeout once you get transcription that fires conversation.end(). clearTimeout when you get the audio-data event.

I'm not a huge fan of implementing a timeout since we don't know much about network traffic, but at least with the events that are fired, you can implement it.

[ppisljar]

i am looking at the way its done in python sample app (where it works perfectly for me)... https://github.com/googlesamples/assistant-sdk-python/blob/master/google-assistant-sdk/googlesamples/assistant/grpc/pushtotalk.py#L109

there is no extra condition anywhere, it just loops over the response packets then ends the conversation, so seems there is a clear notification from the server when the last packet was sent.

looking at the comment in https://github.com/googleapis/googleapis/blob/master/google/assistant/embedded/v1alpha1/embedded_assistant.proto#L255

// This event indicates that the server has detected the end of the user's
    // speech utterance and expects no additional speech. Therefore, the server
    // will not process additional audio (although it may subsequently return
    // additional results). The client should stop sending additional audio
    // data, half-close the gRPC connection, and wait for any additional results
    // until the server closes the gRPC connection.

also seems to suggest that the stream will be closed after last packet is sent ?

[endoplasmic]

Yah, I get a bit lost in the python stuff sadly. Could use a hand getting through that.

As far as your second link, that's actually only to tell you that the user is finished speaking and to not send any more audio to the server. Once you get that message, you'll get the audio (if there is any) right after that.

[ppisljar]

yeah, the function does that, but the comment seems to suggest that the server will close gRPC connection after its done sending data ?

..., and wait for any additional results until the server closes the gRPC connection.

https://developers.google.com/assistant/sdk/reference/library/python/ ... the python library seems to provide more events, one of them being

ON_NO_RESPONSE = 8
The Assistant successfully completed its turn but has nothing to say.

but the same doesn't look to be available in the gRPC api ...