feat: [#47] add tts settings

src/assistant/.env.example

@@ -30,3 +30,10 @@ VOICE_MAX_INPUT_SECONDS=30
 
 OPENAI_API_KEY=sk-1234567890
 OPENAI_STT_MODEL=whisper-1
+
+TTS_YANDEX_API_KEY=
+TTS_YANDEX_AUDIO_FORMAT=oggopus
+TTS_YANDEX_SAMPLE_RATE_HERTZ=48000
+
+TTS_ELEVEN_LABS_API_KEY=
+TTS_ELEVEN_LABS_DEFAULT_VOICE_ID=EXAVITQu4vr4xnSDxMaL

src/assistant/lib/app/settings.py

@@ -22,3 +22,9 @@ class Settings(pydantic_settings.BaseSettings):
 
     proxy: app_split_settings.ProxySettings = pydantic.Field(default_factory=lambda: app_split_settings.ProxySettings())
     voice: app_split_settings.VoiceSettings = pydantic.Field(default_factory=lambda: app_split_settings.VoiceSettings())
+    tts_yandex: app_split_settings.TTSYandexSettings = pydantic.Field(
+        default_factory=lambda: app_split_settings.TTSYandexSettings()
+    )
+    tts_eleven_labs: app_split_settings.TTSElevenLabsSettings = pydantic.Field(
+        default_factory=lambda: app_split_settings.TTSElevenLabsSettings()
+    )

src/assistant/lib/app/split_settings/__init__.py

@@ -5,6 +5,7 @@ from .openai import *
 from .postgres import *
 from .project import *
 from .proxy import *
+from .tts import *
 from .voice import *
 
 __all__ = [
@@ -15,6 +16,8 @@ __all__ = [
     "PostgresSettings",
     "ProjectSettings",
     "ProxySettings",
+    "TTSElevenLabsSettings",
+    "TTSYandexSettings",
     "VoiceSettings",
     "get_logging_config",
 ]

src/assistant/lib/app/split_settings/tts/__init__.py

@@ -0,0 +1,7 @@
+from .eleven_labs import *
+from .yandex import *
+
+__all__ = [
+    "TTSElevenLabsSettings",
+    "TTSYandexSettings",
+]

src/assistant/lib/app/split_settings/tts/eleven_labs.py

@@ -0,0 +1,24 @@
+import pydantic
+import pydantic_settings
+
+import lib.app.split_settings.utils as app_split_settings_utils
+
+
+class TTSElevenLabsSettings(pydantic_settings.BaseSettings):
+    model_config = pydantic_settings.SettingsConfigDict(
+        env_file=app_split_settings_utils.ENV_PATH,
+        env_prefix="TTS_ELEVEN_LABS_",
+        env_file_encoding="utf-8",
+        extra="ignore",
+    )
+
+    api_key: pydantic.SecretStr = pydantic.Field(default=...)
+    default_voice_id: str = "EXAVITQu4vr4xnSDxMaL"
+
+    @property
+    def base_headers(self) -> dict[str, str]:
+        return {
+            "Accept": "audio/mpeg",
+            "Content-Type": "application/json",
+            "xi-api-key": self.api_key.get_secret_value(),
+        }

src/assistant/lib/app/split_settings/tts/yandex.py

@@ -0,0 +1,26 @@
+import typing
+
+import pydantic
+import pydantic_settings
+
+import lib.app.split_settings.utils as app_split_settings_utils
+
+
+class TTSYandexSettings(pydantic_settings.BaseSettings):
+    model_config = pydantic_settings.SettingsConfigDict(
+        env_file=app_split_settings_utils.ENV_PATH,
+        env_prefix="TTS_YANDEX_",
+        env_file_encoding="utf-8",
+        extra="ignore",
+    )
+
+    audio_format: typing.Literal["oggopus", "mp3", "lpcm"] = "oggopus"
+    sample_rate_hertz: int = 48000
+    api_key: pydantic.SecretStr = pydantic.Field(default=...)
+
+    @property
+    def base_headers(self) -> dict[str, str]:
+        return {
+            "Authorization": f"Api-Key {self.api_key.get_secret_value()}",
+            "Content-Type": "application/x-www-form-urlencoded",
+        }