First off, ALL credit for the core functionality of this integration goes to AlexxIT for the original Stream Assist!
I’ve been working on adding functionality to this integration in a fork called Stream Assist CC, and so far I’ve made decent progress! I’m not a developer, so I’ve mostly been trial and erroring my way through this process.
My current issue is that I can’t figure out a way to transition the pipeline from the wake word stage to the STT stage in a way that continues following the original thread, or at least that’s what I think is happening. Basically, I’ve gotten conversation ID preservation, cancellation phrases, and TTS duration calculation all working, but if I manually fire an event to end the wake word, it doesn’t continue transitioning through the rest of the pipeline. If I then add more manual transitions through the rest of the pipeline, I end up losing the conversation ID and cancellation phrase functionality because it seems to be operating outside of the original logic.
At this point it seems like my only option is to reverse engineer the functionality of the pipeline to figure out a way around this, but can anyone here provide more information on how this integration can seamlessly continue as though an actual wake word was used? The code for this is in the init file in the core folder. Any help or guidance would be appreciated!