I’m unsure if it is a bug or not, but I noticed a different parsing of sentences if they are coming from the STT rather then typed in.
I have two examples:
1)
Accendi luce soggiorno (soggiorno is an area and means turn on living room light)
When typed it is parsed correctly and it turns on the light in the living room (soggiorno) area,
When coming from TTS, although identical, it says that there is no device called “luce soggiorno”
2) Accendi luce piccola sala (luce piccola sala is a device)
Typed works correctly.
When from TTS is says “Non c’è nessuna area chiamata piccola sala” (There is no area called “piccola sala”.
It is a bit puzzling, since it is also not consistent. An area is taken for a device in 1 and a device is taken for an area in the second case.
Is this a case where the pattern matching is done in a different order depending on the source of the sentence (keyboard vs. TTS) and it stops before doing the right match?