[jingai]

Also, FWIW, "Alexa, tell Kodi to (play/listen to) boney m." does work here.

I think your primary problem is with this. If we fix that, it will allow you to populate your slots fully, and you would see better matching in general.

[vladchiriloiu]

Hi,

Awesome work. However, it seems I have an issue, I have Python 2.7 installed as per instruction, I have installed the virtual environment. Then I download the kodi-alexa zip, unzip-it on my C drive - other location than Python/Scripts, access it from CMD - ran:
C:\Alexa codes\kodi-alexa>C:\Python27\Scripts\virtualenv venv
C:\Alexa codes\kodi-alexa>venv\Scripts\activate.bat
it gives me:
(venv) C:\Alexa codes\kodi-alexa>
and when I try to run: pip install -r requirements.txt it gives me an error:
Could not open requirements file: [Errno 2] No such file or directory: 'requirements.txt'

Do you know what I am missing? Where I am going wrong? It seems as I am not in the right folder, although in the kodi-alexa folder, there is a requirements.txt file...

Any help is welcomed.
Thank you

[mcl22]

I too believe that the problem is what you explained (and what we already talked about) in point 4. But I also see a problem in point 1. Let me go through them one by one:

1) The history tells me the same what it is written on the cards (is that right?). So it doesn't get me any further because it just states what Alexa did not what she heard (in my opinion). I'm gonna explain it better in point 4.

2) When saying "Alexa, öffne kodi und ..." she opens up the right skill. That works.

3) During all my tests I used the exact utterance to make her do something. The only word I added was "und" which means "and" (open kodi and do something). Also this seems to work as intended.

4) Here we go now. The slot obviously seems to not be generic. But when I tested it with artists I had 300 samples in it. And this is also true for albums, songs and movies. So if the requirement for a slot to be generic is to have around 200 - 300 samples, it doesn't work for me.
As said before the app history doesn't help me either because to me it seems that it doesn't tell me what I really said / Alexa really heard because if eg Boney M. is part of my samples for MUSICARTISTS she performs right. She plays Boney M. and also logs that in the history.
But if I delete Boney M. she plays Bonnie Tyler and also logs that. Just as if I would have said Bonnie Tyler instead of Boney M.. And this is absoloutly reproducible for me. So I don't think it's about my pronounciation or me not speaking clearly enough

5) The skill (code) performs just as it should. I see it in the logs. Whatever it is given it does the right thing (simple match / fuzzy match).

I begin to think that it have something to do with me using the skill in german. I don't know but having those problems with the utterances at first and now with the slots not being generic although they do have enough entries ... I don't know

[jingai]

@mcl22, it's fine to theorize about what might be happening, but I need to verify what's going on, which means I need you to:

1) Ask her to play Boney M,
2) Copy the transcribed text from the Alexa app (or alexa.amazon.com) and paste it here,
3) Copy the skill server log for this request and post it here.

If you could also repeat the above for Bonnie Tyler, that would help too.

Make sure both Bonnie Tyler and Boney M are in your slots for the test.

To answer your question on point 4, your slot should definitely be "generic" if it has 300 items in it. But, it still can be a pronunciation issue. Basically, the way you're saying "boney m" might not be very close to what she expects, but if you add it to the slot, you're weighting her decision toward it. If you take it out of the slot, you're effectively weighting her decision more toward "Bonnie Tyler".

It would be interesting to see what she passes along if you took both out of the slot, but in the end.. I do need to see the transcriptions and logs directly to help further.

[mcl22]

I already did this test and here are the results:

Bonnie Tyler in / Boney M. not:
Trying to play Bonnie Tyler
App: "öffne kodi und spiel musik von bonnie tyler"
Log:
Spiele Musik von bonnie tyler
Searching for artist "bonnie tyler"
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Trying to match: bonnie tyler
Simple match on direct comparison
Located artist "Bonnie Tyler"
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Adding 17 items to the queue...
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***

Trying to play Boney M.
App: "öffne kodi und spiel musik von bonnie tyler"
Log:
Spiele Musik von bonnie tyler
Searching for artist "bonnie tyler"
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Trying to match: bonnie tyler
Simple match on direct comparison
Located artist "Bonnie Tyler"
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Adding 17 items to the queue...
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***

Bonnie Tyler not in / Boney M. not in:
Trying to play Bonnie Tyler
App: "öffne kodi und spiel musik von bonnie tyler"
Log:
Spiele Musik von bonnie tyler
Searching for artist "bonnie tyler"
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Trying to match: bonnie tyler
Simple match on direct comparison
Located artist "Bonnie Tyler"
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Adding 17 items to the queue...
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***
Sending request to http://***/jsonrpc from device amzn1.ask.device.AH***

Trying to play Boney M.
App: "öffne kodi und spiel musik von bonnie tyler"
"öffne kodi und spiel musik von joni james"
"öffne kodi und spiel musik von bon jovi"
"öffne kodi und spiel musik von bangles"
"öffne kodi und spiel musik von rumjacks"
"öffne kodi und spiel musik von boney m."
Boney M. in this case is matched correctly in let's say about 30%. Once it was "none" and the rest is listed above. I understand that I perhaps didn't speak clear enough but those "hits" are all samples and that is what is written in my history. That's why it seems to me that she first tries to match and then writes the history.
The server logs in those cases are as intended.

Bonnie Tyler in / Boney M. in:
Both artists are matched correctly.

Bonnie Tyler not in / Boney M. in:
Boney M. is matched correctly but for Bonnie Tyler Alexa and the log just tells that she couldn't find none. And in one case she instead played Boney M.

So at least in this scenario to me it seems to be saver with both in or both out. But if I only had music from Boney M. and then would get some from Bonnie Tyler it wouldn't work without manually adding her to the samples. And that is only this scenario I accidentaly found. I guess this happens for others, too.

Another problem is that the results above are not absolutly reproducable. Eg after I did all those tests I again deleted both artists from the samples and tried again. This time Alexa either couldn't find none or tried to play Body Count or something else for both artists.

[jingai]

Unfortunately, I don't know what to say at this point. This isn't how Alexa is supposed to work (and it doesn't, for English). It really sounds like your slots are not being converted to "generic," though I don't know why. I suppose it's possible there's a problem with the German version of Alexa at the moment, or that it just behaves differently.. though if the latter is the case, there isn't anything we can do about it. The skill really depends on the literal heard string (not translated) to be passed through in the event that there isn't a solid match with the slot items.

Can any other German users repeat this failure?

[mcl22]

Ja das würde mich jetzt auch mal interessieren, ob es hier noch andere deutsche User gibt, die zumindest ähnlich Probleme haben wie ich. Dann wüßte ich nämlich wenigstens, ob ich vielleicht irgendwie zu blöd bin oder bekäme eine Idee wo man ansetzen kann

Im Augenblick teste ich etliche Szenarien durch und stoße immer wieder auf kleinere oder größere Probleme. Wenn es anderen auch so geht, wäre es echt toll, wenn wir uns da kurzschließen könnten um gemeinsam eine Lösung zu finden. Denn im Augenblick ist es nur ein Rumstochern ohne zu wissen wo man ansetzen könnte.

[vladchiriloiu]

Hi,

Awesome work. However, it seems I have an issue, I have Python 2.7 installed as per instruction, I have installed the virtual environment. Then I download the kodi-alexa zip, unzip-it on my C drive - other location than Python/Scripts, access it from CMD - ran:
C:\Alexa codes\kodi-alexa>C:\Python27\Scripts\virtualenv venv
C:\Alexa codes\kodi-alexa>venv\Scripts\activate.bat
it gives me:
(venv) C:\Alexa codes\kodi-alexa>
and when I try to run: pip install -r requirements.txt it gives me an error:
Could not open requirements file: [Errno 2] No such file or directory: 'requirements.txt'

Do you know what I am missing? Where I am going wrong? It seems as I am not in the right folder, although in the kodi-alexa folder, there is a requirements.txt file...

Any help is welcomed.
Thank you

Any ideas on this issue?

[jingai]

@vladchiriloiu, I don't honestly know. Does this work:

Code:

C:\Alexa codes\kodi-alexa>type requirements.txt

?

[jingai]

@vladchiriloiu, I don't honestly know. Does this work:

Code:

C:\Alexa codes\kodi-alexa>type requirements.txt

?

[vladchiriloiu]

@vladchiriloiu, I don't honestly know. Does this work:

Code:

C:\Alexa codes\kodi-alexa>type requirements.txt

?

Thank you @jimgai, would this create the requirements.txt file? or try to read it? I will try the command when I get home.

[jingai]

It should print the contents of the file.

Sent from my D5803

[jingai]

Kodi-Alexa 2.6.1 has been released, which should hopefully take care of the remaining unicode issues.

If you have unicode characters in your library items that you would like preserved, you should re-generate your slot items again.

[bill_orange]

I have not been able to run the upgrade procedure successfully. I am working in Windows 10. I can start the virtual environment okay.

When I try to run , pip install -r requirements.txt, I get a screen full of 'Requirements already satisfied messages'. I assume that for this upgrade that is normal.

When I try to run 'git pull origin master' or 'git pull https://github.com/m0ngr31/kodi-alexa.git' I get the error:


(venv) C:\Python27>git pull origin master https://github.com/m0ngr31/kodi-alexa.
git
fatal: Not a git repository (or any of the parent directories): .git

or


(venv) C:\Python27\kodi-alexa>git pull origin master https://github.com/m0ngr31/
kodi-alexa.git
fatal: Not a git repository (or any of the parent directories): .git

Is this the wrong syntax for window 10 perhaps? I have not a clue, but I sure don't want to overwrite my .env file. I assume that is the purpose of using git pull origin master vs. git clone.

I tried renaming my folder kodi-alexa, kodi-alexa-master, didn't help.

[Twit]

Yo @bill_orange - It's "git clone" NOT "git pull". I know because I did the same this.

Help File

Cheers

Si