I am doing a standalone version of the components/esphome/assist_satellite.py component in order to learn how the HAVPE works.
I have managed to get a complete pipeline to work (wakeword -> tts) with the default mode of passing a audio url to the VOICE_ASSISTANT_TTS_START event and the HAVPE plays the sound correctly.
However, when I try to stream audio instead (i.e. VOICE_ASSISTANT_TTS_STREAM_START) I can't get it to work. I added some extra debugging to voice_assistant.cpp and noticed that it doesnt enter the "if (this->speaker_ != nullptr)" if clause, meaning the speaker doesnt seem to be available.
Is this intended, i.e. is streaming not supported yet?