jsdio/voice_assistant
Template
1
0
Fork 0
mirror of https://github.com/ijaric/voice_assistant.git synced 2025-05-24 14:33:26 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix: [#28] review fixes

Browse Source
This commit is contained in:
ksieuk 2023-10-11 18:41:10 +03:00
parent 0ef28afa22
commit 48a6f4e167
18 changed files with 258 additions and 227 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
src/assistant/lib/models/__init__.py
View File

@ -3,16 +3,18 @@ from .token import Token
from .tts import * from .tts import *
__all__ = [ __all__ = [
"AVAILABLE_MODELS_TYPE",
"Base", "Base",
"BaseLanguageCodesEnum", "BaseLanguageCodesEnum",
"BaseVoiceModel", "BaseVoiceModel",
"BaseVoiceModel",
"ElevenLabsLanguageCodesEnum", "ElevenLabsLanguageCodesEnum",
"IdCreatedUpdatedBaseMixin", "IdCreatedUpdatedBaseMixin",
"LANGUAGE_CODES_ENUM_TYPE", "LANGUAGE_CODES_ENUM_TYPE",
"LIST_VOICE_MODELS_TYPE",
"TTSCreateRequestModel", "TTSCreateRequestModel",
"TTSCreateResponseModel", "TTSCreateResponseModel",
"TTSSearchVoiceRequestModel", "TTSSearchVoiceRequestModel",
"Token", "Token",
"VoiceModelProvidersEnum",
"YandexLanguageCodesEnum", "YandexLanguageCodesEnum",
] ]

7
src/assistant/lib/models/tts/__init__.py
View File

@ -1,13 +1,16 @@
from .base import * from .models import *
from .languages import * from .voice import *
__all__ = [ __all__ = [
"AVAILABLE_MODELS_TYPE",
"BaseLanguageCodesEnum", "BaseLanguageCodesEnum",
"BaseVoiceModel", "BaseVoiceModel",
"ElevenLabsLanguageCodesEnum", "ElevenLabsLanguageCodesEnum",
"LANGUAGE_CODES_ENUM_TYPE", "LANGUAGE_CODES_ENUM_TYPE",
"LIST_VOICE_MODELS_TYPE",
"TTSCreateRequestModel", "TTSCreateRequestModel",
"TTSCreateResponseModel", "TTSCreateResponseModel",
"TTSSearchVoiceRequestModel", "TTSSearchVoiceRequestModel",
"VoiceModelProvidersEnum",
"YandexLanguageCodesEnum", "YandexLanguageCodesEnum",
] ]

26
src/assistant/lib/models/tts/base.py → src/assistant/lib/models/tts/models.py
View File

@ -1,14 +1,16 @@
import typing
import pydantic import pydantic
import lib.models.tts.languages as models_tts_languages import lib.models.tts.voice as models_tts_voice
import lib.models.tts.voice.languages as models_tts_languages
AVAILABLE_MODELS_TYPE = models_tts_voice.YandexVoiceModel | models_tts_voice.ElevenLabsVoiceModel
LIST_VOICE_MODELS_TYPE = models_tts_voice.YandexListVoiceModelsModel | models_tts_voice.ElevenLabsListVoiceModelsModel
class TTSCreateRequestModel(pydantic.BaseModel): class TTSCreateRequestModel(pydantic.BaseModel):
model_config = pydantic.ConfigDict(use_enum_values=True) model_config = pydantic.ConfigDict(use_enum_values=True)
voice_model_name: str voice_model: AVAILABLE_MODELS_TYPE
text: str text: str
@ -16,22 +18,6 @@ class TTSCreateResponseModel(pydantic.BaseModel):
audio_content: bytes audio_content: bytes
class BaseVoiceModel(pydantic.BaseModel):
voice_id: str
voice_name: str | None = None
languages: list[models_tts_languages.LANGUAGE_CODES_ENUM_TYPE]
company_name: str
@pydantic.model_validator(mode="before")
@classmethod
def check_voice_name_exists(cls, data: typing.Any) -> typing.Any:
voice_id = data.get("voice_id")
voice_name = data.get("voice_name")
if not voice_name and voice_id:
data["voice_name"] = voice_id
return data
class TTSSearchVoiceRequestModel(pydantic.BaseModel): class TTSSearchVoiceRequestModel(pydantic.BaseModel):
voice_id: str | None = None voice_id: str | None = None
voice_name: str | None = None voice_name: str | None = None

17
src/assistant/lib/models/tts/voice/__init__.py Normal file
View File

@ -0,0 +1,17 @@
from .base import *
from .eleven_labs import *
from .languages import *
from .yandex import *
__all__ = [
"BaseLanguageCodesEnum",
"BaseVoiceModel",
"ElevenLabsLanguageCodesEnum",
"ElevenLabsListVoiceModelsModel",
"ElevenLabsVoiceModel",
"LANGUAGE_CODES_ENUM_TYPE",
"VoiceModelProvidersEnum",
"YandexLanguageCodesEnum",
"YandexListVoiceModelsModel",
"YandexVoiceModel",
]

27
src/assistant/lib/models/tts/voice/base.py Normal file
View File

@ -0,0 +1,27 @@
import enum
import typing
import pydantic
import lib.models.tts.voice.languages as models_tts_languages
class VoiceModelProvidersEnum(enum.Enum):
YANDEX = "yandex"
ELEVEN_LABS = "eleven_labs"
class BaseVoiceModel(pydantic.BaseModel):
voice_id: str
voice_name: str | None = None
languages: list[models_tts_languages.LANGUAGE_CODES_ENUM_TYPE]
provider: VoiceModelProvidersEnum
@pydantic.model_validator(mode="before")
@classmethod
def check_voice_name_exists(cls, data: typing.Any) -> typing.Any:
voice_id = data.get("voice_id")
voice_name = data.get("voice_name")
if not voice_name and voice_id:
data["voice_name"] = voice_id
return data

75
src/assistant/lib/models/tts/voice/eleven_labs.py Normal file
View File

@ -0,0 +1,75 @@
import typing
import pydantic
import lib.models.tts.voice.base as models_tts_base
import lib.models.tts.voice.languages as models_tts_languages
class ElevenLabsVoiceModel(models_tts_base.BaseVoiceModel):
model_config = pydantic.ConfigDict(use_enum_values=True)
voice_id: str
voice_name: str | None = None
languages: list[models_tts_languages.LANGUAGE_CODES_ENUM_TYPE]
provider: models_tts_base.VoiceModelProvidersEnum = models_tts_base.VoiceModelProvidersEnum.ELEVEN_LABS
class ElevenLabsListVoiceModelsModel(pydantic.BaseModel):
models: list[ElevenLabsVoiceModel] = [
ElevenLabsVoiceModel(
voice_id="eleven_multilingual_v1",
languages=[
models_tts_languages.ElevenLabsLanguageCodesEnum.ENGLISH,
models_tts_languages.ElevenLabsLanguageCodesEnum.GERMAN,
models_tts_languages.ElevenLabsLanguageCodesEnum.POLISH,
models_tts_languages.ElevenLabsLanguageCodesEnum.SPANISH,
models_tts_languages.ElevenLabsLanguageCodesEnum.ITALIAN,
models_tts_languages.ElevenLabsLanguageCodesEnum.FRENCH,
models_tts_languages.ElevenLabsLanguageCodesEnum.PORTUGUESE,
models_tts_languages.ElevenLabsLanguageCodesEnum.HINDI,
models_tts_languages.ElevenLabsLanguageCodesEnum.ARABIC,
],
),
ElevenLabsVoiceModel(
voice_id="eleven_multilingual_v2",
languages=[
models_tts_languages.ElevenLabsLanguageCodesEnum.ENGLISH,
models_tts_languages.ElevenLabsLanguageCodesEnum.JAPANESE,
models_tts_languages.ElevenLabsLanguageCodesEnum.CHINESE,
models_tts_languages.ElevenLabsLanguageCodesEnum.GERMAN,
models_tts_languages.ElevenLabsLanguageCodesEnum.HINDI,
models_tts_languages.ElevenLabsLanguageCodesEnum.FRENCH,
models_tts_languages.ElevenLabsLanguageCodesEnum.KOREAN,
models_tts_languages.ElevenLabsLanguageCodesEnum.PORTUGUESE,
models_tts_languages.ElevenLabsLanguageCodesEnum.ITALIAN,
models_tts_languages.ElevenLabsLanguageCodesEnum.SPANISH,
models_tts_languages.ElevenLabsLanguageCodesEnum.INDONESIAN,
models_tts_languages.ElevenLabsLanguageCodesEnum.DUTCH,
models_tts_languages.ElevenLabsLanguageCodesEnum.TURKISH,
models_tts_languages.ElevenLabsLanguageCodesEnum.FILIPINO,
models_tts_languages.ElevenLabsLanguageCodesEnum.POLISH,
models_tts_languages.ElevenLabsLanguageCodesEnum.SWEDISH,
models_tts_languages.ElevenLabsLanguageCodesEnum.BULGARIAN,
models_tts_languages.ElevenLabsLanguageCodesEnum.ROMANIAN,
models_tts_languages.ElevenLabsLanguageCodesEnum.ARABIC,
models_tts_languages.ElevenLabsLanguageCodesEnum.CZECH,
models_tts_languages.ElevenLabsLanguageCodesEnum.GREEK,
models_tts_languages.ElevenLabsLanguageCodesEnum.FINNISH,
models_tts_languages.ElevenLabsLanguageCodesEnum.CROATIAN,
models_tts_languages.ElevenLabsLanguageCodesEnum.MALAY,
models_tts_languages.ElevenLabsLanguageCodesEnum.SLOVAK,
models_tts_languages.ElevenLabsLanguageCodesEnum.DANISH,
models_tts_languages.ElevenLabsLanguageCodesEnum.TAMIL,
models_tts_languages.ElevenLabsLanguageCodesEnum.UKRAINIAN,
],
),
ElevenLabsVoiceModel(
voice_id="eleven_multilingual_v2",
languages=[models_tts_languages.ElevenLabsLanguageCodesEnum.ENGLISH],
),
]
@classmethod
def from_api(cls, voice_models_from_api: list[dict[str, typing.Any]]) -> typing.Self:
voice_models = [ElevenLabsVoiceModel.model_validate(voice_model) for voice_model in voice_models_from_api]
return ElevenLabsListVoiceModelsModel(models=voice_models)

0
src/assistant/lib/models/tts/languages.py → src/assistant/lib/models/tts/voice/languages.py
View File

97
src/assistant/lib/models/tts/voice/yandex.py Normal file
View File

@ -0,0 +1,97 @@
import typing
import pydantic
import lib.models.tts.voice.base as models_tts_base
import lib.models.tts.voice.languages as models_tts_languages
class YandexVoiceModel(models_tts_base.BaseVoiceModel):
voice_id: str
voice_name: str | None = None
languages: list[models_tts_languages.LANGUAGE_CODES_ENUM_TYPE]
provider: models_tts_base.VoiceModelProvidersEnum = models_tts_base.VoiceModelProvidersEnum.YANDEX
role: str | None = None
@pydantic.model_validator(mode="before")
@classmethod
def check_voice_name_exists(cls, data: typing.Any) -> typing.Any:
voice_id = data.get("voice_id")
voice_name = data.get("voice_name")
role = data.get("role")
if not voice_name and voice_id:
data["voice_name"] = f"{voice_id} {role}" if role else voice_id
return data
class YandexListVoiceModelsModel(pydantic.BaseModel):
models: list[YandexVoiceModel] = [
YandexVoiceModel(
voice_id="ermil", role="neutral", languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]
),
YandexVoiceModel(
voice_id="ermil", role="good", languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]
),
YandexVoiceModel(
voice_id="alena", role="neutral", languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]
),
YandexVoiceModel(
voice_id="alena", role="good", languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]
),
YandexVoiceModel(
voice_id="jane", role="neutral", languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]
),
YandexVoiceModel(
voice_id="jane", role="good", languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]
),
YandexVoiceModel(
voice_id="jane", role="evil", languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]
),
YandexVoiceModel(
voice_id="omazh", role="neutral", languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]
),
YandexVoiceModel(
voice_id="omazh", role="evil", languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]
),
YandexVoiceModel(
voice_id="zahar", role="neutral", languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]
),
YandexVoiceModel(
voice_id="zahar", role="good", languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]
),
YandexVoiceModel(
voice_id="filipp", role=None, languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]
),
YandexVoiceModel(
voice_id="madirus", role=None, languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]
),
YandexVoiceModel(voice_id="dasha", role=None, languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="julia", role=None, languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="lera", role=None, languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(
voice_id="marina", role=None, languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]
),
YandexVoiceModel(
voice_id="alexander", role=None, languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]
),
YandexVoiceModel(
voice_id="kirill", role=None, languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]
),
YandexVoiceModel(voice_id="anton", role=None, languages=[models_tts_languages.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="john", role=None, languages=[models_tts_languages.YandexLanguageCodesEnum.ENGLISH]),
YandexVoiceModel(voice_id="amira", role=None, languages=[models_tts_languages.YandexLanguageCodesEnum.KAZAKH]),
YandexVoiceModel(voice_id="madi", role=None, languages=[models_tts_languages.YandexLanguageCodesEnum.KAZAKH]),
YandexVoiceModel(voice_id="lea", role=None, languages=[models_tts_languages.YandexLanguageCodesEnum.GERMAN]),
YandexVoiceModel(
voice_id="naomi", role="modern", languages=[models_tts_languages.YandexLanguageCodesEnum.HEBREW]
),
YandexVoiceModel(
voice_id="naomi", role="classic", languages=[models_tts_languages.YandexLanguageCodesEnum.HEBREW]
),
YandexVoiceModel(voice_id="nigora", role=None, languages=[models_tts_languages.YandexLanguageCodesEnum.UZBEK]),
]
@classmethod
def from_api(cls, voice_models_from_api: list[dict[str, typing.Any]]) -> typing.Self:
voice_models = [YandexVoiceModel.model_validate(voice_model) for voice_model in voice_models_from_api]
return YandexListVoiceModelsModel(models=voice_models)

14
src/assistant/lib/tts/models/__init__.py
View File

@ -1,17 +1,5 @@
from .repository import * from .protocols import *
from .service import *
from .utils import *
from .voice import *
__all__ = [ __all__ = [
"ElevenLabsListVoiceModelsModel",
"ElevenLabsVoiceModel",
"LIST_VOICE_MODELS_TYPE",
"LIST_VOICE_MODELS_TYPE",
"TTSCreateRequestServiceModel",
"TTSCreateResponseServiceModel",
"TTSRepositoryProtocol", "TTSRepositoryProtocol",
"VOICE_MODELS_TYPE",
"YandexListVoiceModelsModel",
"YandexVoiceModel",
] ]

14
src/assistant/lib/tts/models/protocols.py Normal file
View File

@ -0,0 +1,14 @@
import typing
import lib.models as models
class TTSRepositoryProtocol(typing.Protocol):
def get_audio_as_bytes(self, request: models.TTSCreateRequestModel) -> models.TTSCreateResponseModel:
...
def get_voice_model_by_name(self, voice_model_name: str) -> models.BaseVoiceModel | None:
...
def get_voice_models_by_fields(self, fields: models.TTSSearchVoiceRequestModel) -> models.LIST_VOICE_MODELS_TYPE:
...

17
src/assistant/lib/tts/models/repository.py
View File

@ -1,17 +0,0 @@
import typing
import lib.models as models
import lib.tts.models.utils as tts_models_utils
class TTSRepositoryProtocol(typing.Protocol):
def get_audio_as_bytes_from_text(self, text: str) -> models.TTSCreateResponseModel:
...
def get_voice_model_by_name(self, voice_model_name: str) -> models.BaseVoiceModel | None:
...
def get_voice_models_by_fields(
self, fields: models.TTSSearchVoiceRequestModel
) -> tts_models_utils.LIST_VOICE_MODELS_TYPE:
...

9
src/assistant/lib/tts/models/service.py
View File

@ -1,9 +0,0 @@
import lib.models as models
class TTSCreateRequestServiceModel(models.TTSCreateRequestModel):
...
class TTSCreateResponseServiceModel(models.TTSCreateResponseModel):
...

4
src/assistant/lib/tts/models/utils.py
View File

@ -1,4 +0,0 @@
import lib.tts.models.voice as tts_models_voice
VOICE_MODELS_TYPE = tts_models_voice.YandexVoiceModel | tts_models_voice.ElevenLabsVoiceModel
LIST_VOICE_MODELS_TYPE = tts_models_voice.YandexListVoiceModelsModel | tts_models_voice.ElevenLabsListVoiceModelsModel

9
src/assistant/lib/tts/models/voice/__init__.py
View File

@ -1,9 +0,0 @@
from .eleven_labs import *
from .yandex import *
__all__ = [
"ElevenLabsListVoiceModelsModel",
"ElevenLabsVoiceModel",
"YandexListVoiceModelsModel",
"YandexVoiceModel",
]

73
src/assistant/lib/tts/models/voice/eleven_labs.py
View File

@ -1,73 +0,0 @@
import typing
import pydantic
import lib.models as models
class ElevenLabsVoiceModel(models.BaseVoiceModel):
voice_id: str
voice_name: str | None = None
languages: list[models.LANGUAGE_CODES_ENUM_TYPE]
company_name: str = "eleven labs"
class ElevenLabsListVoiceModelsModel(pydantic.BaseModel):
models: list[ElevenLabsVoiceModel] = [
ElevenLabsVoiceModel(
voice_id="eleven_multilingual_v1",
languages=[
models.ElevenLabsLanguageCodesEnum.ENGLISH,
models.ElevenLabsLanguageCodesEnum.GERMAN,
models.ElevenLabsLanguageCodesEnum.POLISH,
models.ElevenLabsLanguageCodesEnum.SPANISH,
models.ElevenLabsLanguageCodesEnum.ITALIAN,
models.ElevenLabsLanguageCodesEnum.FRENCH,
models.ElevenLabsLanguageCodesEnum.PORTUGUESE,
models.ElevenLabsLanguageCodesEnum.HINDI,
models.ElevenLabsLanguageCodesEnum.ARABIC,
],
),
ElevenLabsVoiceModel(
voice_id="eleven_multilingual_v2",
languages=[
models.ElevenLabsLanguageCodesEnum.ENGLISH,
models.ElevenLabsLanguageCodesEnum.JAPANESE,
models.ElevenLabsLanguageCodesEnum.CHINESE,
models.ElevenLabsLanguageCodesEnum.GERMAN,
models.ElevenLabsLanguageCodesEnum.HINDI,
models.ElevenLabsLanguageCodesEnum.FRENCH,
models.ElevenLabsLanguageCodesEnum.KOREAN,
models.ElevenLabsLanguageCodesEnum.PORTUGUESE,
models.ElevenLabsLanguageCodesEnum.ITALIAN,
models.ElevenLabsLanguageCodesEnum.SPANISH,
models.ElevenLabsLanguageCodesEnum.INDONESIAN,
models.ElevenLabsLanguageCodesEnum.DUTCH,
models.ElevenLabsLanguageCodesEnum.TURKISH,
models.ElevenLabsLanguageCodesEnum.FILIPINO,
models.ElevenLabsLanguageCodesEnum.POLISH,
models.ElevenLabsLanguageCodesEnum.SWEDISH,
models.ElevenLabsLanguageCodesEnum.BULGARIAN,
models.ElevenLabsLanguageCodesEnum.ROMANIAN,
models.ElevenLabsLanguageCodesEnum.ARABIC,
models.ElevenLabsLanguageCodesEnum.CZECH,
models.ElevenLabsLanguageCodesEnum.GREEK,
models.ElevenLabsLanguageCodesEnum.FINNISH,
models.ElevenLabsLanguageCodesEnum.CROATIAN,
models.ElevenLabsLanguageCodesEnum.MALAY,
models.ElevenLabsLanguageCodesEnum.SLOVAK,
models.ElevenLabsLanguageCodesEnum.DANISH,
models.ElevenLabsLanguageCodesEnum.TAMIL,
models.ElevenLabsLanguageCodesEnum.UKRAINIAN,
],
),
ElevenLabsVoiceModel(
voice_id="eleven_multilingual_v2",
languages=[models.ElevenLabsLanguageCodesEnum.ENGLISH],
),
]
@classmethod
def from_api(cls, voice_models_from_api: list[dict[str, typing.Any]]) -> typing.Self:
voice_models = [ElevenLabsVoiceModel.model_validate(voice_model) for voice_model in voice_models_from_api]
return ElevenLabsListVoiceModelsModel(models=voice_models)

60
src/assistant/lib/tts/models/voice/yandex.py
View File

@ -1,60 +0,0 @@
import typing
import pydantic
import lib.models as models
class YandexVoiceModel(models.BaseVoiceModel):
voice_id: str
voice_name: str | None = None
languages: list[models.LANGUAGE_CODES_ENUM_TYPE]
company_name: str = "yandex"
role: str | None = None
@pydantic.model_validator(mode="before")
@classmethod
def check_voice_name_exists(cls, data: typing.Any) -> typing.Any:
voice_id = data.get("voice_id")
voice_name = data.get("voice_name")
role = data.get("role")
if not voice_name and voice_id:
data["voice_name"] = f"{voice_id} {role}" if role else voice_id
return data
class YandexListVoiceModelsModel(pydantic.BaseModel):
models: list[YandexVoiceModel] = [
YandexVoiceModel(voice_id="ermil", role="neutral", languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="ermil", role="good", languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="alena", role="neutral", languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="alena", role="good", languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="jane", role="neutral", languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="jane", role="good", languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="jane", role="evil", languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="omazh", role="neutral", languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="omazh", role="evil", languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="zahar", role="neutral", languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="zahar", role="good", languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="filipp", role=None, languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="madirus", role=None, languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="dasha", role=None, languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="julia", role=None, languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="lera", role=None, languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="marina", role=None, languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="alexander", role=None, languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="kirill", role=None, languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="anton", role=None, languages=[models.YandexLanguageCodesEnum.RUSSIAN]),
YandexVoiceModel(voice_id="john", role=None, languages=[models.YandexLanguageCodesEnum.ENGLISH]),
YandexVoiceModel(voice_id="amira", role=None, languages=[models.YandexLanguageCodesEnum.KAZAKH]),
YandexVoiceModel(voice_id="madi", role=None, languages=[models.YandexLanguageCodesEnum.KAZAKH]),
YandexVoiceModel(voice_id="lea", role=None, languages=[models.YandexLanguageCodesEnum.GERMAN]),
YandexVoiceModel(voice_id="naomi", role="modern", languages=[models.YandexLanguageCodesEnum.HEBREW]),
YandexVoiceModel(voice_id="naomi", role="classic", languages=[models.YandexLanguageCodesEnum.HEBREW]),
YandexVoiceModel(voice_id="nigora", role=None, languages=[models.YandexLanguageCodesEnum.UZBEK]),
]
@classmethod
def from_api(cls, voice_models_from_api: list[dict[str, typing.Any]]) -> typing.Self:
voice_models = [YandexVoiceModel.model_validate(voice_model) for voice_model in voice_models_from_api]
return YandexListVoiceModelsModel(models=voice_models)

8
src/assistant/lib/tts/repositories/base.py
View File

@ -1,7 +1,6 @@
import abc import abc
import lib.models as models import lib.models as models
import lib.tts.models as tts_models
class HttpClient: # Mocked class todo remove and use real http client from lib.clients.http_client class HttpClient: # Mocked class todo remove and use real http client from lib.clients.http_client
@ -15,11 +14,11 @@ class TTSBaseRepository(abc.ABC):
@property @property
@abc.abstractmethod @abc.abstractmethod
def voice_models(self) -> tts_models.LIST_VOICE_MODELS_TYPE: def voice_models(self) -> models.LIST_VOICE_MODELS_TYPE:
... ...
@abc.abstractmethod @abc.abstractmethod
def get_audio_as_bytes_from_text(self, text: str) -> models.TTSCreateResponseModel: def get_audio_as_bytes(self, request: models.TTSCreateRequestModel) -> models.TTSCreateResponseModel:
raise NotImplementedError raise NotImplementedError
def get_voice_model_by_name(self, voice_model_name: str) -> models.BaseVoiceModel | None: def get_voice_model_by_name(self, voice_model_name: str) -> models.BaseVoiceModel | None:
@ -34,7 +33,7 @@ class TTSBaseRepository(abc.ABC):
def get_list_voice_models_by_fields( def get_list_voice_models_by_fields(
self, fields: models.TTSSearchVoiceRequestModel self, fields: models.TTSSearchVoiceRequestModel
) -> list[tts_models.VOICE_MODELS_TYPE]: ) -> list[models.AVAILABLE_MODELS_TYPE]:
""" """
Search voice model by fields Search voice model by fields
:param fields: Any fields from TTSSearchVoiceRequestModel :param fields: Any fields from TTSSearchVoiceRequestModel
@ -43,7 +42,6 @@ class TTSBaseRepository(abc.ABC):
fields_dump = fields.model_dump(exclude_none=True) fields_dump = fields.model_dump(exclude_none=True)
voice_models_response = [] voice_models_response = []
for voice_model in self.voice_models.models: for voice_model in self.voice_models.models:
voice_model: tts_models.VOICE_MODELS_TYPE
for field, field_value in fields_dump.items(): for field, field_value in fields_dump.items():
if field == "languages": # language is a list if field == "languages": # language is a list
language_names: set[str] = {item.name for item in field_value} language_names: set[str] = {item.name for item in field_value}

24
src/assistant/lib/tts/services.py
View File

@ -7,32 +7,28 @@ class TTSService:
def __init__( def __init__(
self, self,
settings: app_settings.Settings, settings: app_settings.Settings,
tts_repositories: list[tts_models.TTSRepositoryProtocol], repositories: dict[models.VoiceModelProvidersEnum, tts_models.TTSRepositoryProtocol],
): ):
self.settings = settings self.settings = settings
self.tts_repositories = tts_repositories self.repositories = repositories
def get_audio_as_bytes_from_text(self, tts_request: models.TTSCreateRequestModel) -> models.TTSCreateResponseModel: def get_audio_as_bytes(self, request: models.TTSCreateRequestModel) -> models.TTSCreateResponseModel:
for repository in self.tts_repositories: model = request.voice_model
voice_model = repository.get_voice_model_by_name(tts_request.voice_model_name) repository = self.repositories[model.provider]
if voice_model: audio_response = repository.get_audio_as_bytes(request)
audio_response = repository.get_audio_as_bytes_from_text(tts_request.text)
break
else:
raise ValueError(f"Voice model {tts_request.voice_model_name} not found")
return audio_response return audio_response
def get_voice_model_by_name(self, voice_model_name: str) -> models.BaseVoiceModel | None: def get_voice_model_by_name(self, voice_model_name: str) -> models.BaseVoiceModel | None:
for repository in self.tts_repositories: for repository in self.repositories.values():
voice_model = repository.get_voice_model_by_name(voice_model_name) voice_model = repository.get_voice_model_by_name(voice_model_name)
if voice_model: if voice_model:
return voice_model return voice_model
def get_list_voice_models_by_fields( def get_list_voice_models_by_fields(
self, fields: models.TTSSearchVoiceRequestModel self, fields: models.TTSSearchVoiceRequestModel
) -> list[tts_models.VOICE_MODELS_TYPE]: ) -> list[models.AVAILABLE_MODELS_TYPE]:
response_models: list[tts_models.VOICE_MODELS_TYPE] = [] response_models: list[models.AVAILABLE_MODELS_TYPE] = []
for repository in self.tts_repositories: for repository in self.repositories.values():
voice_models = repository.get_voice_models_by_fields(fields) voice_models = repository.get_voice_models_by_fields(fields)
if voice_models.models: if voice_models.models:
response_models.extend(voice_models.models) response_models.extend(voice_models.models)