Build an open future with us.

Invest in Mycroft and become a community partner.

Testing and Feedback for Telegram Skill


#21

This is really cool. I currently have Home Assistant send me notifications via Telegram when a door opens and when the Nest detects motion. Very useful.

However, as I’m sure we’re all trying to become a little less cloud-y around here, I’d like to float the idea of an XMPP skill. That way it would be Open Source without caveats, and users wouldn’t be dependent on the servers of a single company which [cough] is of interest to governments and law enforcement.

But that’s neither here nor there. Excellent work. Looks very useful. Might have to try it out.


#22

@Luke I have your telegram skill installed and it shows in home.mycroft - have also saved the token there, etc - however, when I message the bot I get nothing - no Chat ID to enter, no response whatsoever… any ideas where to start?


#23

@Luke has been MIA since August. Hopefully he will chime in. The skill was working well up until recently.


#24

Hi, will try to look into this soon.


#25

Sorry for my long silence, did a new commit.
hopefully the skill should now work as expected.


#26

Great that you like it :slight_smile:


#27

Did a new commit which should fix your issue, if not please give me quick Feedback.


#28

sorry that it took me so long to answear you.

is your mark1 up to date?
did you try to install it as a custom skill (put the GitHub link under custom skill on home.mycroft.ai)

or did you manage to get it installed already?


#29

Welcome back. Change was pushed to my device in the past 1/2 hour and all is working as expected. Thanks, this is definitely one of my primary skills.


#30

Thank you for your feedback and i’m really glad to hear that you like the skill :slight_smile:


#31

fixed and improved it a little.

My Dev-Machine is a Ubuntu VM, does anyone have a Mark I at hand to test if everything works as expected?

Changes:

  1. New function sendMycroftSay: it just takes a message and puts “say” in front of it and sends it to the message bus as an utterance (why use the say skill if you could send it directly to the message bus?: if the say-skill get’s an update -> telegram-skill does not need to be modified and the phrase “say SOMETHING” will not likely change in the future)
  2. The skill should not only send you a message when loaded, mycroft should also say “Telegram skill loaded”.
  3. Muting should be more reliable (it happened before, that only some part of the sentence was muted)
  4. If Mute is enabled (at home.mycroft.ai) but does not work (because of problems with alsaaudio), the skill now says out loud “There is a Problem with alsa Audio, mute is not working” when skill is loaded
  5. If skill was not able to compare the Mycroft Device Name with the Name you stated on home.mycroft.ai, mycroft will say out loud “No or incorrect Device Name specified! Your DeviceName is: YOURUNITNAME”

ToDo:
Multi-Sentence answears from mycroft are not yet supported by the telegram skill.
for example:
User: Hey Mycroft, what can you do?
Mycroft: i can tell you the time, what the weather is like……. (first sentence -> is sent back to user via telegram, mute is working)
Mycroft: i have 33 Skills installed …… (second sentence -> is not sent to user, no mute)

best would be if it was possible to send an utterance with context-tag (something like telegram-message) and everything mycroft does/say because of this utterance get’s the same context-tag
If anyone has an idea how to achieve this, i would appreciate it.


#32

So, The magic mirror skill @dmwilsonkc , and my kodi skill have the ability to display utterances as notification callbacks. This may work for multiline responses. In the kodi skill when you say turn kodi notifications on this redirects all responses textually to the kodi interface (or magic mirror). I suspect this may work for you. Also i can confirm that your skill works on picroft.


#33

Thanks for the hint.

if i am not mistaken you have an Event that triggers on every utterance, that was something i thought about, but found it pretty resource intensive.

For example if you have kodi and telegram skill installed and there is an utterance, both skills would be called because of this event.

but i will give it a try, maybe i have an epiphany while programming :slight_smile: