• 1
  • 171
  • 172
  • 173(current)
  • 174
  • 175
  • 176
Release Kanzi: Amazon Alexa skill for Kodi
Hey,

I think I got pretty far, Kanzi skill deployed and all, visible in my Alexa skills. But when I try to use a Kanzi command I get the error from Alexa "Request access violation, some long string amzn.1.............." do you have any idea what I did wrong in the process? Does the forwarded port has to be 8080 or are other ports allowed aswell? I added the Amazon app ID to my config file aswell but no change.

Thanks in advance.
Reply
For thouse who got "configuration file not valid" message, at validating point, check special chars in Kodi password at kodi.config file.

I'm getting this error when deploying

D:\Proyectos\Lexigram>lexigram deploy koko
 √ Check config file
 √ Validate config file
 √ Check skill directory
 √ Remove old repo code
 √ Download latest source code
 √ Extract source code
 √ Extract source code - Part 2
 √ Update skill data
 √ Update slot values
 × Deploy skill
   → Error deploying. Please try again

And at 'errorlog.txt'
Call update-skill error.
Error code: 404
{
  "message": "Resource not found."
}


Any clue?

Many thnanks.

EDIT: it seems my problem was because of not fill the skill id at adding the Alexa trigger (it default selects to check it). Now i success deploying koko, and could deploy again filling the right Skill id (until first deploy i havent it) Now i got an error deploying kanzi, but it works.
EDIT2: today I successfully deploy kanzi after second attempt, first with a bussy network error, and second success. And I haven't changed anything, just login again and launch deploy.
Reply
(2020-04-26, 19:33)Andrajos Wrote: For thouse who got "configuration file not valid" message, at validating point, check special chars in Kodi password at kodi.config file.

I'm getting this error when deploying

D:\Proyectos\Lexigram>lexigram deploy koko
 √ Check config file
 √ Validate config file
 √ Check skill directory
 √ Remove old repo code
 √ Download latest source code
 √ Extract source code
 √ Extract source code - Part 2
 √ Update skill data
 √ Update slot values
 × Deploy skill
   → Error deploying. Please try again

And at 'errorlog.txt'
Call update-skill error.
Error code: 404
{
  "message": "Resource not found."
}


Any clue?

Many thnanks.
Do you have your kodi device on?
That was it for me the first time, it got stuck because it couldn't connect to it.
Reply
(2020-04-30, 19:51)cryonic90 Wrote:
(2020-04-26, 19:33)Andrajos Wrote: For thouse who got "configuration file not valid" message, at validating point, check special chars in Kodi password at kodi.config file.

I'm getting this error when deploying

D:\Proyectos\Lexigram>lexigram deploy koko
 √ Check config file
 √ Validate config file
 √ Check skill directory
 √ Remove old repo code
 √ Download latest source code
 √ Extract source code
 √ Extract source code - Part 2
 √ Update skill data
 √ Update slot values
 × Deploy skill
   → Error deploying. Please try again

And at 'errorlog.txt'
Call update-skill error.
Error code: 404
{
  "message": "Resource not found."
}


Any clue?

Many thnanks.
Do you have your kodi device on?
That was it for me the first time, it got stuck because it couldn't connect to it. 
Yes, if not, or if the password is wrong, it fails at step 

√ Update slot values
Reply
I've done the deploy. Went though all looked good.

I enabled the Kanzi Developer skill in my Alexa phone app.

When I ask Kanzi to do something I get an activity showing in the Alexa phone app like:

Protocol Violation
Kanzi
Request Identifier: amzn1.echo-api.request.5d1cf98c-21b4-4ae9-9356-f686f54f78fa
The application response did not conform to the specification. Please ensure that you are using the latest client library and that your responses comply with the Application Service Interface.
Alexa heard: "alexa ask kanzi to play bad boys"
Voice feedback

How do I troubleshoot this?

Thanks I would very much appreciate some assistance on how to go about resolving this problem.

My kodi.config is:


# Global configuration
#
[global]
# Currently supported languages: en, de, fr, it
language = en

# By default, if you ask to play media without specifying the type, the skill
# will search through the entire library to find a match.
#
# For example, if you ask, "Alexa, ask Kodi to play ninety nine red balloons,"
# the skill will search library items in the following order to find the song:
#
#   Movies -> Shows -> Artists -> Songs -> Albums
#
# If your library is really large, this can take some time to complete.  While
# it will eventually execute, Alexa may report that the skill timed out.  If
# this behavior bothers you, you can disable deep searches.
#
# When disabled, Alexa will simply provide help in the response directing you
# to issue a play command that includes the media type of the item.
deep_search = yes

# Limit number of items we add to playlists.
#
# Some requests will cause the skill to create a temporary playlist in Kodi
# before executing your command.  This operation is expensive in Kodi, so we
# provide the option here to limit the number of items added.
#
# The default is 100 to match Kodi's own 'party mode.'  Higher values will
# make these operations slower, but if your setup can handle it, you're free
# to set this to whatever you like.
#
# Set to an empty string for unlimited.
playlist_max_items = 100

# Search result limits for new (unwatched) Movies and Episodes.
#
# By default we ask kodi to return a max of 100 results per JSON request.
# Here you can set this figure to optimize the system for you.
# Results are always retrieved in order of newest episodes and movies first.
#
# Lower figures will produce faster results and limit them to only media that
# you've recently added to the library.
unwatched_shows_max_results     = 100
unwatched_episodes_max_results  = 100
unwatched_movies_max_results    = 100

# Set logging level.  Possible values are:
#
# CRITICAL
# ERROR
# WARNING
# INFO
# DEBUG
loglevel = INFO

# Log sensitive or personally identifying information.
#
# Disabling this prevents the skill from logging the target address for Kodi
# and your device IDs.
#
# It is enabled by default to make it easy for you to identify your Echo
# devices for mapping to specific instances of Kodi.
logsensitive = yes


# Global Alexa skill configuration
#
[alexa]
# Set skill_id to enable verification of requests by Flask-Ask.
skill_id =

# Maximum number of items to generate per slot.
#
# If the skill is failing to save, you may need to reduce this.  Conversely,
# if Alexa is having problems distinguishing your requests (executing the
# wrong Intent, for instance), you may need to increase this.
slot_items_max = 100


#
# Default parameters common to all Alexa devices and the web simulator
#
[DEFAULT]
# The Kodi webserver only supports HTTP, but if you've set up a reverse HTTPS
# proxy you can change this to https.
scheme   = http
address  = ausvette1.theworkpc.com
port     = 8080
# If using a reverse proxy you might need to add an extra bit to the url
# before "jsonrpc" (don't use slashes before or after).
subpath  =
username = kodi
password = ********

# Caching of Kodi responses.
#
# Provide an Amazon S3 bucket or a directory name here and provide credentials
# for one of the cache backends below to enable.
cache_bucket = ausvette

# Amazon S3 cache backend.
#
# Provide an AWS user key and key ID here to enable.
s3_cache_aws_access_key_id =  AKIA4BTMG7QWR4ENTFEZ
s3_cache_aws_secret_access_key =  q+xXqxIAvKUj1B+NxDSpdLmJrcYxTiDZzG96EGjo

# ownCloud/nextCloud cache backend.
#
# Provide the base URL for your ownCloud server and a valid user and password
# combination to enable.
owncloud_cache_url =
owncloud_cache_user =
owncloud_cache_password =

# Read timeout -- how long to wait for responses from Kodi before giving up.
#
# Normally there is no need to change this.
read_timeout = 120

# For some commands, we don't care about the response from Kodi, so we send
# them "fire-and-forget."  If you've got a proxy server in the middle, though,
# it's possible the command might not make it to Kodi before the skill exits.
#
# If Alexa is telling you she has completed a command but it sometimes does
# not actually execute on Kodi, you may need to increase this.
read_timeout_async = 0.01

# Set shutdown to 'quit' if you'd like "Alexa, tell Kodi to shut down"
# to quit Kodi instead of shutting down the system.
shutdown = quit

# Your local time zone for responses that include absolute times.
# See https://en.wikipedia.org/wiki/List_of_tz...time_zones
#
# For example, if you are in the Eastern US time zone, you would use:
# timezone = US/Eastern
#
# Leave empty if you don't need or want absolute time responses.  An example
# is asking when the currently playing item will end.  If you have this
# defined, it will also tell you the wall-clock time when the item will
# conclude.
timezone = Australia/Queensland

# Setup MongoDB to handle playing music from Kodi to your Alexa device
# Must accept agreement saying I'm not liable for stolen information
# since your username and password for Kodi will be stored in plaintext
# in a database and will be transferred over the internet to a HTTPS
# proxy server for you to have this functionality.
#
# mongodb_uri will look like: `mongodb://user:[email protected]:port/database`
# If you're using alternative proxy from Heroku, alt_proxty looks like:
# `https://my-kodi-proxy.herokuapp.com/proxy?file=`
mongodb_uri=
use_proxy=
alt_proxy=
accept_music_warning=


# Override default values by specifying them below for individual devices.
#
# The usual reason to do this is to associate a given Echo device with a
# particular instance of Kodi in your house, but you can override any of
# the other configuration variables from the DEFAULT section here too.
#
# Device IDs can be found in the skill server / lambda log.
Reply
When I use the Alexa Simulator on the Alexa Developer Console - typing in the manual commands gives " The requested skill did not provide a valid response"

How do I trouble shoot this? I haven't been able to find any logs for this.

Also is the url for the skills server the address for the herokuapp.com server - something like https://xxxx.herokuapp.com?

Thanks!
Reply
Can anyone who has this working please post their json for the intent section? The json and utterance samples are in a format that now isn't valid in the dev interface so it's really beefing me out.

Thanks in advance
Matt
Reply
(2020-05-09, 20:35)1981mattwilson Wrote: Can anyone who has this working please post their json for the intent section? The json and utterance samples are in a format that now isn't valid in the dev interface so it's really beefing me out.

Thanks in advance
Matt
If you tell me where to find it, I can post it.
Reply
(2020-05-09, 20:35)1981mattwilson Wrote: Can anyone who has this working please post their json for the intent section? The json and utterance samples are in a format that now isn't valid in the dev interface so it's really beefing me out.

Thanks in advance
Matt

Is Kanzi still working?

I am not sure if I should keep troubleshooting what I have setup, or if it just doesn't work now?
Reply
(2020-05-22, 09:37)ausvette Wrote:
(2020-05-09, 20:35)1981mattwilson Wrote: Can anyone who has this working please post their json for the intent section? The json and utterance samples are in a format that now isn't valid in the dev interface so it's really beefing me out.

Thanks in advance
Matt

Is Kanzi still working?

I am not sure if I should keep troubleshooting what I have setup, or if it just doesn't work now? 

It works for me, but not fully (some artist and albums doesn't recognize)
Reply
I can't get Kanzi to deploy fully when using the docker container. I have it working with nginx for reverse proxy setup. I can access my kanzi server which I used docker for. I get a 405 error when hitting that. Here are the log excerpts. Has anyone resolved this issue before?

::: 2020-06-04 19:21:21 :::
Error: Command failed: /usr/lib/node_modules/lexigram-cli/dist/../node_modules/.bin/ask deploy -t model
[Error]: Invalid response for interactionModel status from get-skill-status request.

[18:54:54] → Error deploying. Please try again,
[18:54:51] → Deploying skill slot data and intents.,
[18:54:54] Deploy skill [failed],
[18:54:49] Deploy skill [started],
[18:54:49] → Deploying skill information.,
    Kanzi 01.24.2020-dXXXXXab successfully deployed with invocation name kanzi at https://kanzi.XXXXXX.com/kanzi/
Reply
(2018-05-03, 00:56)m0ngr31 Wrote: Yeah, it looked like it finished with that part, just got stuck on uploading the model. I'm not sure what the issue is, but there is a way to figure it out.

Go to your Skill page in Amazon, and browse to the JSON editor:
Image

Then copy everything from the kanzi/models/en-US.json file into there and hit save. That should tell us what's going on.
This solved my problem! I used the en-AU.json file and it will now do some of my bidding. Thanks for all the dev work you have done here. Still not sure why my model fails to deploy through the cli though.
Reply
(2020-04-12, 05:34)serv Wrote: It fails to deploy the Kanzi skill with the self-hosted linuxserver docker. Fails on Lambda as well. Failed on a clean build new server with npm. Tried rebuilding several times and have ensured the endpoint is reachable.

In the error log I'm seeing this https://paste.kodi.tv/igahugirod I truncated the last bit as it repeats for a while.
Error arises from /usr/lib/node_modules/lexigram-cli/node_modules/ask-cli/lib/deploy/deploy-model.js on line 189.

Alexa developer console simple states that "The model build has failed with an unspecified error. This could be from an invalid model or we could be experiencing intermittent issues. Please try again."

Shows 10 good full builds and one failed for each deploy attempt. I assume one language is the hangup. It loaded 137 intents and 15 slot types, but does show 9 Utterance Conflicts in EN-US (back, go back, go home, go to sleep, hold, home, navigate home, right, shut down).

Don't know if it's because I'm running a very large library, because alexa wants a two word invocation, or more likely some other bug.

Fixed issue with following steps:
html:
Removed references to all other languages besides EN-US and EN_GB in:
 /node_modules/lexigram-cli/action.ts
 /node_modules/lexigram-cli/InteractionModel.ts
 /node_modules/lexigram-cli/kanzi-skill.json
 /node_modules/lexigram-cli/koko-skill.json
 /node_modules/lexigram-cli/dist/actions.js
 /node_modules/lexigram-cli/dist/InteractionModel.js
 /kanzi/skill.json
 deleted all but en-US and en-GB from /kanzi/models/
 
 in the generated lambda.zip:
    deleted all languages but templates.en.yaml sample_slotvals.en.txt and utterances.en.txt
    /alexa.py

Edited node_modules/lexigram-cli/actions.ts and node_modules/lexigram-cli/dist/actions.ts to stop it from removing old repo code, downloading new source, and extracting new source
Copied contents of the lambda.zip to /home/kanzi/kanzi/source/repo/

-I have exact same problem and got the same error. 9 Utterance Conflicts. Seems some kindow bug to me.
Reply
Atleast I got it working. I got always Utterance Conflicts every language what I tryed. The fix was for me, that  I set alexa location to

settings -> device settings -> my alexa -> Time zone Easter daylight time (new york) and language English united kingtom.

In alexa dev console I set language to English (US) in the left right. Then Interaction Model -> JSON Editor and pasted there en-US.json. It is located in /app/kanzi/kanzi/models/ in docker setup.
Reply
My Kanzi skill has been deploying for over an hour now. Anyone have any idea how to resolve?

https://imgur.com/12xwOv1
Reply
  • 1
  • 171
  • 172
  • 173(current)
  • 174
  • 175
  • 176

Logout Mark Read Team Forum Stats Members Help
Kanzi: Amazon Alexa skill for Kodi15
This forum uses Lukasz Tkacz MyBB addons.