tts_websocketserver

TTS WebsocketServer

This is a websocket server based on rgws for integration of our inference pipeline with the production team
The server exposes an RPC method called process_text with the following I/O:
Input:

{"text": "<A SENTENCE>"}

Output:

{
    'velocity': <int>,
    'pitch': <int>,
    'source': <str>,
    'qualities': [<str>, <str>, ...],
    'latent_sample': [<float>, <float>, <float>, ...]
}

Setup

pip install - r requirements.txt

if this fails for any reason, run the following commands:

pip install rgws
pip install git+https://[email protected]/TheSoundOfAIOSR/rg_text_to_sound.git#"subdirectory=tts_pipeline"

Then run the following:

cd src
python -m tts_websocketserver.initialize

Test run

Server:

cd src
python -m tts_websocketserver.tts_server

Client:

cd src
python -m tts_websocketserver.simple_client