Issue with weather in Italian

[denics]

when I ask the temperature in italian: che tempo fa? I get for example: Attualmente è nuvoloso e 25.2 gradi centigradi the issue is with the number: 25.2 is read as 252 instead of "25 gradi e 2. A suggestion to simplify temperature reading would be a 0.5 approximation, in order to have "25 gradi" or "26 gradi" and maybe, have a new commandqual'è la temperatura precisa, in English:what is the precise temperature`` to skip the approximation.

[emphasize]

I don't think this is a skill issue. What is converting 25.2 into 252 ?

the temperature specifically gets no special formatting besides the dialog renderer (ie format the dialog string with the number). Be aware that you also could ask for fahrenheit scale. So you would have to stick with "25 punto 2 gradi" nevertheless.

One degree is too much of a leeway for certain applications/informational needs I'm hesitant to add complexity in form of (an/multiple) additional intent(s).

[denics]

then the issue should be raised to core, @JarbasAl do you get the same with Portuguese?

[emphasize]

I'm not quite clear about the problem. does it voice out https://translate.google.de/?sl=de&tl=it&text=es%20ist%20252%20grad&op=translate instead of https://translate.google.de/?sl=de&tl=it&text=es%20ist%2025.2%20grad&op=translate

which tts is involved?

[denics]

The issue is how the number is read. In the logs the answer is correct: 25.2, but then it is spell as 252. I am testing on ovos-docker, with this configuration:

{
  "lang": "it-it",
  "tts": {
    "module": "ovos-tts-plugin-mimic3-server",
    "fallback_module": "",
    "ovos-tts-plugin-mimic3-server": {
        "voice": "it_IT/mls_low"
    }
  },
  "stt": {
    "module": "ovos-stt-plugin-server",
    "fallback_module": "ovos-stt-plugin-vosk",
    "ovos-stt-plugin-vosk": {
        "model": "https://alphacephei.com/vosk/models/vosk-model-small-it-0.22.zip"
    }
  }
}

[emphasize]

so, it's ultimately a mimic3-it synth issue

[JarbasAl]

seems you need to change TTS then, mimic3 wont be getting more updates....

[denics]

Ok, I'll test with piper. Should I open a new issue for documenting this on the mimic3 repo?

[JarbasAl]

Ok, I'll test with piper. Should I open a new issue for documenting this on the mimic3 repo?

this would be model/voice specific, any issues should go into the original mycroft mimic3 repo, but don't expect any answers in there either other than maybe someone telling you to move to piper

i'd recommend just using piper which has all the same mimic3 voices, and then open issues in piper repo if needed, I'm sure Piper wants (and gives you) feedback and is actively improving the voices!