mycroftai / fallback-unknown Goto Github PK

The number of utterances available in the en-us don't have great variety, which leads to Mycroft sounding very repetitive when he doesn't understand what you're saying.

This can be fixed by adding more options to the unknown.dialog file, which I intend to do momentarily.

I'm translating via pootle the skill. There are some expressions I think cannot be translated into spanish properly. I see other romance languages like Italian where are translated, but I think they did wrong because romance languages need to put the verb in the proper tense, while germanic/english ones have the auxiliar particle for that purpose.
So "where did" can be translated literally to "dónde hizo", but that really would mean "where he did". Besides, and that is the real problem, "where did she go?" is translated to "¿dónde fue?", "where did he born?" to "¿dónde nació?" and so on.
The same happens with the rest of auxiliars: do, does, did, will, shall, etc

Bug Description
Anytime the "QuestionsAnsweringSkill" (I am assuming that is the same as this skill?) is triggered with an unknown utterance, it times out. This is then strangely followed up with a response from the timer skill: "how long of a timer" which is then repeated once (if user does not provide a follow up utterance).

To Reproduce
Steps to reproduce the behavior:

1. Say an unknown (intent) utterance
2. Mycroft will search
3. it will time out
4. this will trigger the timer "how long of a timer" will be repeated twice

Expected behavior
When an intent cannot be determined, the fallback will handle it, not time out, and not trigger the timer.

Log files
Here is an actual example from the logs where I said "what version" instead of "what Mycroft version" as an utterance. But as stated, it will do this for every single unknown intent.

QuestionsAnswersSkill | Searching for what version 12:27:12.238 | INFO | 19423 | QuestionsAnswersSkill | Timeout occured check responses 12:27:14.962 | INFO | 19429 | __main__:handle_record_begin:41 | Begin Recording... 12:27:17.985 | INFO | 19429 | __main__:handle_record_end:49 | End Recording... 12:27:19.596 | ERROR | 19429 | mycroft.client.speech.listener:transcribe:239 | list index out of range 12:27:19.599 | ERROR | 19429 | mycroft.client.speech.listener:transcribe:240 | Speech Recognition could not understand audio 12:27:22.041 | INFO | 19429 | __main__:handle_record_begin:41 | Begin Recording... 12:27:25.061 | INFO | 19429 | __main__:handle_record_end:49 | End Recording... 12:27:26.863 | ERROR | 19429 | mycroft.client.speech.listener:transcribe:239 | list index out of range 12:27:26.865 | ERROR | 19429 | mycroft.client.speech.listener:transcribe:240 | Speech Recognition could not understand audio

Environment

• Device type: (running in a docker container) with a raspberrypi 4.
• OS: Ubuntu container (Rasbian OS aarch64 on the pi itself)
• Mycroft-core version: 21.02

Additional context
I have uninstalled this skill and re-installed it, I found a temporary fix to the timer trigger by simply removing the timer (although it seems to auto-install itself by default again and the new install of the timer skill doesn't seem to fix this issue either), I have completely re-installed Mycroft by deploying a new instance of the docker container, and also updated Mycroft within the container. Nonetheless, this issue persists.

Thank you very much for your time!