Charmed Quark Systems, Ltd. - Support Forums and Community
New Amazon Echo Support - Printable Version

+- Charmed Quark Systems, Ltd. - Support Forums and Community (https://www.charmedquark.com/vb_forums)
+-- Forum: General Discussion (https://www.charmedquark.com/vb_forums/forumdisplay.php?fid=3)
+--- Forum: CQC Support (https://www.charmedquark.com/vb_forums/forumdisplay.php?fid=9)
+--- Thread: New Amazon Echo Support (/showthread.php?tid=9423)



New Amazon Echo Support - IVB - 02-19-2016

jkmonroe Wrote:you're right that it should be consistent, but who's to say that there can't be two?

Smile

Apologies I wasn't clear.

I'm not saying CQC shouldn't offer two, only that I wouldn't personally take advantage of more than one at a time. I also think I'd pick "Alexa, ask Jarvis to ..." as thats faster than speak-waitforresponse-speak-waitforresult. But as mentioned, I also acknowledge that my family is far from the "normal" path.


New Amazon Echo Support - Dean Roddey - 02-19-2016

You can do an ongoing conversation. However, it requires something on the other end maintain context across those invocations, because they aren't one continuous thing. They go to the Echo, to CQC, then back every time. That makes things considerably more complex so I didn't try to support it.

That's maybe something better supported by some built in assistant that we provided or something, since we could much more easily maintain that context.

One thing I suggested to them via their survey thingie is the ability to have each Echo be named and have that name come through, so that you can know which one it came from. Currently I don't think there's a good way to do that. That way you could take a command without any need to indicate location and automatically do what's right for the room is was invoked from.


New Amazon Echo Support - IVB - 02-19-2016

Dean Roddey Wrote:One thing I suggested to them via their survey thingie is the ability to have each Echo be named and have that name come through, so that you can know which one it came from. Currently I don't think there's a good way to do that. That way you could take a command without any need to indicate location and automatically do what's right for the room is was invoked from.

ok now THAT is interesting...


New Amazon Echo Support - batwater - 04-08-2016

Dean is there a way to configure this to pass through the text spoken to the Echo without processing? E.g. can I just have one entry in the local CQC configuration that accepts anything?


New Amazon Echo Support - Dean Roddey - 04-08-2016

I'm not sure what you are meaning there? As for the actual text spoken, we never see that. We only see the digested version, the intent and the parameters.


New Amazon Echo Support - batwater - 04-08-2016

The EchoTasker Skill passes the entire message to the target endpoint for processing. The relevant portion of the intents file is below:

Code:
"intents": [
    {
      "intent":  "TaskerIntent",
      "slots": [
        {
            "name": "Message",
            "type": "AMAZON.LITERAL"
        }
      ]
    },

For utterances it uses this:

TaskerIntent {hello|Message}
TaskerIntent {what's up|Message}
TaskerIntent {i am home|Message}
TaskerIntent {what are you doing|Message}
TaskerIntent {who ate my captain crunch|Message}
TaskerIntent {what is taking you so long|Message}
TaskerIntent {i forgot to pick up groceries today|Message}
TaskerIntent {The bees don't want us to eat honey|Message}

Not sure why it works at this point, all I know is that regardless of what I ask Echo to send the text in its entirety arrives at the endpoint.

I would like to have the raw text arrive in CQC for processing and not have to deal with the mess setting up a plethora of utterances.


New Amazon Echo Support - Dean Roddey - 04-08-2016

Oh, OK, that works because they are setting up the entire message as a parameter. That probably won't work for us because it would conflict with existing intents, i.e. it would make them all ambiguous.

You don't really have to set up any utterances. Just use the ones that are there in the sample info. They aren't literal, they are just patterns for matching spoken commands to intents. And they will likely save you a lot of work, because they allow you to trigger a single command based on a number of variations of the same thing.


New Amazon Echo Support - batwater - 04-08-2016

Not sure that answers my "can I get the raw message in CQC" question, maybe I'm just being dense and not getting what you are saying. I'm wanting to process this in a different way but have the 2 way (response back) communication with the Echo.

If I'm starting from scratch I want to put in the bare minimum (single intent if possible) to get the full message through.


New Amazon Echo Support - Dean Roddey - 04-08-2016

You can't do that. The web server understands those intents that are defined by by the setup info we provide. It won't know what to do with any other type of intent. You'll have to work within the existing scheme. It works fine and it's actually pretty convenient, because it lets you map multiple variations of a target thing to a single action or macro.

If you want to do it differently, you'd have to write your own stuff on the Amazon side and on your network side, using perhaps the XML GW interface to talk to CQC.


New Amazon Echo Support - batwater - 04-08-2016

Okay understand. Let me ask a different way, can I funnel what I'm sending from Amazon into a single already defined intent and mapped macro handler on the CQC side and have the full text come through or does something have to match between the 2? How about if on the skill handler side I pre-pend a trigger word to the text sent to CQC so that the Macro fires based on a "match?" Will I be able to get to the full string of the text sent from the echo in CQCActParm somehow?

What is there is pretty convenient but still requires a strict mapping structure, I want to try and avoid the initial mapping structure constraints and use a processing module for the text that was designed specifically for text (speech) processing. The end result back to CQC is a tokenized message just like you are creating. It's a different way of skinning the same cat