Motivation for this change

Mozilla TTS is a text to speech engine based on pytorch and tensorflow. @Mic92 and I have been successfully using it for the last few days in combination with https://rhasspy.readthedocs.io, which we might eventually package, not sure yet.

Demo: https://shells.darmstadt.ccc.de/~hexa/nixpkgs.wav

To get this running one needs to download a TTS and Vocoder model, check out https://github.com/mozilla/TTS/wiki/Released-Models, from where we pulled a commit, that matched a given version.

  systemd.services.tts = {
    after = [ "network.target" ];
    wantedBy = [ "multi-user.target" ];
    script = ''
      ${pkgs.tts}/bin/tts-server \
        --vocoder_config ./vocoder/config.json \
        --vocoder_checkpoint ./vocoder/checkpoint_1450000.pth.tar \
        --tts_config ./tts/config.json \
        --tts_checkpoint ./tts/checkpoint_130000.pth.tar
    '';
    serviceConfig = {
      User = "youruser";
      # needed for pulseaudio
      Environment = "XDG_RUNTIME_DIR=/run/user/1000";
      WorkingDirectory = "/path/to/models";
    };
  };

Then reproduce the demo like this:

curl http://localhost:5002/api/tts\\?text\\=Dear%20NixOS%20Maintainer,%20I%20am%20Mozilla%20TTS.%20I%20am%20ready%20to%20be%20merged,%20please%20review%20me\\. > /tmp/nixpkgs.wav

Things done

• Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS linux)
• Built on platform(s)
  • NixOS
  • macOS
  • other Linux distributions
• Tested via one or more NixOS test(s) if existing and applicable for the change (look inside nixos/tests)
• Tested compilation of all pkgs that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review wip"
• Tested execution of all binary files (usually in ./result/bin/)
• Determined the impact on package closure size (by running nix path-info -S before and after)
• Ensured that relevant documentation is up to date
• Fits CONTRIBUTING.md.