Support for TTS via Hermes/MQTT

[nanosonde]

It would be nice if rhasspy could also use Hermes protocol for speaking text.

Actual use-case for me would be the usage of AWS Polly for TTS. The Python script below listens for Hermes TTS "say" command plays the returned "WAV" data via Hermes Audio Server protocol: https://github.com/hcooper/snips-tts-polly

• subscribes to hermes/tts/say to pick up new TTS requests.
• sends converted audio to hermes/audioServer/default/playBytes.
• closes Snips session via hermes/tts/sayFinished.

The above script also caches already generated WAVs.

[synesthesiam]

Not sure if this is exactly what you mean, but Rhasspy will handle the playBytes message if you set the audio output to Hermes.

[nanosonde]

Somehow I am still attracted by the idea of having each step of the (Snips) processing pipeline available via MQTT(Hermes). For now I could reuse the snips-dialogue-manager which required the „low-level“ Hermes TTS API until a replacement was found.

But you are right:intent handling should not be part of rhasspy which my idea would imply.