tts.py

Go to the documentation of this file.

"""Support for the Microsoft Cognitive Services text-to-speech service."""
 
import logging
 
from pycsspeechtts import pycsspeechtts
from requests.exceptions import HTTPError
import voluptuous as vol
 
from homeassistant.components.tts import (
    CONF_LANG,
    PLATFORM_SCHEMA as TTS_PLATFORM_SCHEMA,
    Provider,
)
from homeassistant.const import CONF_API_KEY, CONF_REGION, CONF_TYPE, PERCENTAGE
from homeassistant.generated.microsoft_tts import SUPPORTED_LANGUAGES
import homeassistant.helpers.config_validation as cv
 
CONF_GENDER = "gender"
CONF_OUTPUT = "output"
CONF_RATE = "rate"
CONF_VOLUME = "volume"
CONF_PITCH = "pitch"
CONF_CONTOUR = "contour"
_LOGGER = logging.getLogger(__name__)
 
GENDERS = ["Female", "Male"]
 
DEFAULT_LANG = "en-us"
DEFAULT_GENDER = "Female"
DEFAULT_TYPE = "JennyNeural"
DEFAULT_OUTPUT = "audio-24khz-96kbitrate-mono-mp3"
DEFAULT_RATE = 0
DEFAULT_VOLUME = 0
DEFAULT_PITCH = "default"
DEFAULT_CONTOUR = ""
DEFAULT_REGION = "eastus"
 
PLATFORM_SCHEMA = TTS_PLATFORM_SCHEMA.extend(
    {
        vol.Required(CONF_API_KEY): cv.string,
        vol.Optional(CONF_LANG, default=DEFAULT_LANG): vol.In(SUPPORTED_LANGUAGES),
        vol.Optional(CONF_GENDER, default=DEFAULT_GENDER): vol.In(GENDERS),
        vol.Optional(CONF_TYPE, default=DEFAULT_TYPE): cv.string,
        vol.Optional(CONF_RATE, default=DEFAULT_RATE): vol.All(
            vol.Coerce(int), vol.Range(-100, 100)
        ),
        vol.Optional(CONF_VOLUME, default=DEFAULT_VOLUME): vol.All(
            vol.Coerce(int), vol.Range(-100, 100)
        ),
        vol.Optional(CONF_PITCH, default=DEFAULT_PITCH): cv.string,
        vol.Optional(CONF_CONTOUR, default=DEFAULT_CONTOUR): cv.string,
        vol.Optional(CONF_REGION, default=DEFAULT_REGION): cv.string,
    }
)
 
 
def get_engine(hass, config, discovery_info=None):
    """Set up Microsoft speech component."""
    return MicrosoftProvider(
        config[CONF_API_KEY],
        config[CONF_LANG],
        config[CONF_GENDER],
        config[CONF_TYPE],
        config[CONF_RATE],
        config[CONF_VOLUME],
        config[CONF_PITCH],
        config[CONF_CONTOUR],
        config[CONF_REGION],
    )
 
 
class MicrosoftProvider(Provider):
    """The Microsoft speech API provider."""
 
    def __init__(
        self, apikey, lang, gender, ttype, rate, volume, pitch, contour, region
    ):
        """Init Microsoft TTS service."""
        self._apikey_apikey = apikey
        self._lang_lang = lang
        self._gender_gender = gender
        self._type_type = ttype
        self._output_output = DEFAULT_OUTPUT
        self._rate_rate = f"{rate}{PERCENTAGE}"
        self._volume_volume = f"{volume}{PERCENTAGE}"
        self._pitch_pitch = pitch
        self._contour_contour = contour
        self._region_region = region
        self.namename = "Microsoft"
 
    @property
    def default_language(self):
        """Return the default language."""
        return self._lang_lang
 
    @property
    def supported_languages(self):
        """Return list of supported languages."""
        return SUPPORTED_LANGUAGES
 
    @property
    def supported_options(self):
        """Return list of supported options like voice, emotion."""
        return [CONF_GENDER, CONF_TYPE]
 
    @property
    def default_options(self):
        """Return a dict include default options."""
        return {CONF_GENDER: self._gender_gender, CONF_TYPE: self._type_type}
 
    def get_tts_audio(self, message, language, options):
        """Load TTS from Microsoft."""
        if language is None:
            language = self._lang_lang
 
        try:
            trans = pycsspeechtts.TTSTranslator(self._apikey_apikey, self._region_region)
            data = trans.speak(
                language=language,
                gender=options[CONF_GENDER],
                voiceType=options[CONF_TYPE],
                output=self._output_output,
                rate=self._rate_rate,
                volume=self._volume_volume,
                pitch=self._pitch_pitch,
                contour=self._contour_contour,
                text=message,
            )
        except HTTPError as ex:
            _LOGGER.error("Error occurred for Microsoft TTS: %s", ex)
            return (None, None)
        return ("mp3", data)