Port TTS

[DEATHB4DEFEAT]

Description

https://github.com/SpaceStationUA/station/pull/7 https://github.com/SpaceStationUA/station/pull/8 I need some help finding a bunch of English voices we can link as defaults.

This uses Piper to generate the audio. Windows support is untested, but exists. There is very little setup required; install Piper, find some models and make their prototypes (or download the "defaults" to bin/Content.Server/data/tts/models and rename them accordingly), set tts.enabled to true, maybe change tts.model_path (starting the path with data/ will put it in the server's datadir), and you're good to go.

---

TODO

• [x] IC language support
• [x] Fix volume
• [ ] Improve sanitization
• [x] Cache messages
• [x] Automatically delete messages
• [ ] Queue audio
• [ ] Review the code
• [x] Figure out what SSML is

---

Media

https://github.com/user-attachments/assets/e65457bd-6ea7-4bf7-9d9c-5be1f6a80fbb

---

Changelog

:cl:

• add: Added TTS

[Evgencheg]

@DEATHB4DEFEAT i think better to cherry-pick is - corvax tts that works with GPU but needs an external hosting

[bruhmogus]

please tell me this will have a server/client config to disable

[Rxup]

@DEATHB4DEFEAT https://github.com/Rxup/ss14-tts-api it's support random voice generator via tts_test.ipynb and can use https://models.silero.ai/models/tts/en/v3_en.pt change in https://github.com/Rxup/ss14-tts-api/blob/main/ss14tts.py#L31

[github-actions[bot]]

This pull request has conflicts, please resolve those before we can evaluate the pull request.

[Crono209ggg]

Bro this is so cool!