[sveni_lee]

ahh... okay thank you...
so far as I can see PVR is not included yet... no Problem, will give a try...

[bill_orange]

I see version 2.5 is out. I have a copy of version 2.1 up and running well. As the ver 2.5 readme points out, the setup is a lot different, especially the virtual environment. Rather than try to update my installation, would it make better sense to create a new Kodi skill following the ver 2.5 readme with a different name and invocation (new skill on my Lambda site, new skill in the Developer's area and new folder structure on my local machine) and delete the ver 2.1 stuff once 2.5 is up and running.

Are there any consequences to this approach that I am mot seeing? Would this be considerably harder than an upgrade?

Bill

[jingai]

@jingai,
thanks for your work on this project.

yesterday I started to chang the wsgi.py for German languages but know I see I would be better to do this within the templates.de.yaml.

I start also to include all changes from the freemans13 fork for PVR. But if you say that version 2.5 will come soon and will include this I could spend my time with translation.

v2.5 will support German utterances (phrases you speak to Alexa), but we need someone to translate her responses (templates.de.yaml).

v2.5 will not include PVR support, but it's on the TODO list.

[jingai]

I'm really looking forward for the pvr function. But concerning your answer for my sts, 2pac, etc. problem I wonder what is going to happen with the pvr feature. The coolest thing would be to ask kodi just for the channel name. But most channels (in Germany) do have names that, i guess, would cause the same problems. Examples: RTL, Sat1, Pro7, ARD, ZDF, n-tv, n24, ...
It's not possible to do some kind of linking, is it? I mean a possibility to tell the skill something like "if you get a s.t.s. then try sts" or "R.T.L. is RTL" or "Sat eins ist Sat1".

We will figure something out.

I'have already tried version 2.5 and the problem with the Umlaute is still there. If I ask kodi to play "Die Ärzte" the response is "Spiele die Arzte", also in 2.5. Better and really funny example is: "play music from Abstürzende Brieftauben" - response "spiele Absturz Ende Brieftauben"

For everyone: Version 2.5 is not released yet -- the flask-ask branch is under active development. Not yelling, just making it bold so people see it.

As for the umlauts, please raise an Issue on github and we'll see what we can do to address it.

Concerning the custom slots: I don't have a problem with including everything on the first run. But if alexa only gets the right match if the asked movie, artist whatever is part of the list, I'm gonna run into problems every time I add a new movie or artist.

Add everything you can, then leave it alone. The skill uses fuzzy matching in addition to the slots -- it's not either/or. It will still most likely work, and if you encounter a title it doesn't work with, raise an Issue for it on github and add it to your slots at least temporarily to make it work.

Not that that happens every day. But updating the slots every time I add something new wouldn't be that comfortable.

You only need to update slots if the skill has trouble recognizing something you've added.

Last question: is there some kind of tutorial for running the skill on my own webserver? I do already run a webserver on a raspberry. So the ressource is there.

It's just WSGI + HTTPS and for 2.5, the entry-point is alexa.wsgi. There is no tutorial, because there are plenty of resources out there for setting up a web server. Plus, this isn't really something I personally want to provide tech support for, honestly.

[jingai]

I see version 2.5 is out.

Version 2.5 is not released yet. You're free to utilize our development branch, but be aware that it will probably break (it's currently broken, even).

I have a copy of version 2.1 up and running well. As the ver 2.5 readme points out, the setup is a lot different, especially the virtual environment. Rather than try to update my installation, would it make better sense to create a new Kodi skill following the ver 2.5 readme with a different name and invocation (new skill on my Lambda site, new skill in the Developer's area and new folder structure on my local machine) and delete the ver 2.1 stuff once 2.5 is up and running.

Are there any consequences to this approach that I am mot seeing? Would this be considerably harder than an upgrade?

Please wait until we've ironed out the details. The upgrade procedure isn't yet complete.

[bill_orange]

Got it! I will wait. Since 2.1 works fine, I don't see any reason to make trouble for myself.

[mcl22]

For everyone: Version 2.5 is not released yet -- the flask-ask branch is under active development. Not yelling, just making it bold so people see it.

...

It's just WSGI + HTTPS and for 2.5, the entry-point is alexa.wsgi. There is no tutorial, because there are plenty of resources out there for setting up a web server. Plus, this isn't really something I personally want to provide tech support for, honestly.

Didn't want to upset you
I know how to set up a webserver. Only the python wsgi stuff was new to me. Meanwhile I have it up and running. I see an improvement in responsetime. But I think I made the wrong decision when I used my raspberry 2 for this test. I see it's cpu going up to 100% every time I use alexa. Perhaps I'm gonna try it with my raspberry 3. May I ask you what hardware you use for your webserver?

[sveni_lee]

may be you could explain how to setup a Web-Server in this case.


Gesendet von iPhone mit Tapatalk Pro

[jingai]

My webserver is on an actual PC so I wouldn't see that kind of resource usage.

And I'm not upset at all -- just wanted to clear up that 2.5 is not out and isn't ready for general use.

[jingai]

I'm sorry, but I'm just too busy to explain how to set up a Web server, especially as I know I'll get a bunch of follow up questions.

[mcl22]

Ok, I was just wondering because the server on the pi works so far and the requests are much faster than it was before. Nevertheless alexa often seems to not get the response in time. So I say something like "Alexa ask kodi to play Terminator" and the movie is played after, let's say 2-3 seconds, but alexa tells me that she couldn't get a response in time.

Whatever, I'm gonna try it with my raspberry 3. Perhaps this is gonna improve response time even more.

[jingai]

The timeout on responses is hard-coded by Amazon to 10 seconds. Nothing we can do about that, but it should still go through.

[digiltd]

Ok, I was just wondering because the server on the pi works so far and the requests are much faster than it was before. Nevertheless alexa often seems to not get the response in time. So I say something like "Alexa ask kodi to play Terminator" and the movie is played after, let's say 2-3 seconds, but alexa tells me that she couldn't get a response in time.

Whatever, I'm gonna try it with my raspberry 3. Perhaps this is gonna improve response time even more.

Just a thought, do you get the same delay if you say " Alexa, ask kodi to play the movie Terminator"?

or for music "Alexa, play the song blahblah" and "Alexa, play the album blahblah

[mcl22]

Music takes longer. The timeouts mostly occure when asking for some music like "play artist ..."

[jingai]

It'll depend on what type of media you have more of, as the skill has to request the entirety of a given media type from Kodi to do its matching.