Allow to specify TTS language in the service call. (#5047)
* Allow to specify TTS language in the service call. * Allow to specify TTS language in the service call. * Respect 79 char limit * Fix "Too many blank lines" * Fix "Too many blank lines" * Fix "Too many blank lines" * Change language to be optional parameter of *get_tts_audio * Change language to be optional parameter of *get_tts_audio * Respect 79 char limit * Don't pass "None * Use default of "None" for TTS language * Use default of "None" for TTS language * Don't pass "None" * Change TTS cache key to be hash_lang_engine * Change language from demo to en * Fix wrong replace
This commit is contained in:
1 parent
68865ec27b
commit
4728fa8da6
8 files changed
+206
-38
No files matched your search
@@ -5,8 +5,9 @@ For more details about this component, please refer to the documentation at
|
||||
https://home-assistant.io/components/tts/
|
||||
"""
|
||||
import asyncio
|
||||
import logging
|
||||
import functools
|
||||
import hashlib
|
||||
import logging
|
||||
import mimetypes
|
||||
import os
|
||||
import re
|
||||
@@ -48,8 +49,10 @@ SERVICE_CLEAR_CACHE = 'clear_cache'
|
||||
|
||||
ATTR_MESSAGE = 'message'
|
||||
ATTR_CACHE = 'cache'
|
||||
ATTR_LANGUAGE = 'language'
|
||||
|
||||
_RE_VOICE_FILE = re.compile(r"([a-f0-9]{40})_([a-z]+)\.[a-z0-9]{3,4}")
|
||||
_RE_VOICE_FILE = re.compile(r"([a-f0-9]{40})_([^_]+)_([a-z]+)\.[a-z0-9]{3,4}")
|
||||
KEY_PATTERN = '{}_{}_{}'
|
||||
|
||||
PLATFORM_SCHEMA = cv.PLATFORM_SCHEMA.extend({
|
||||
vol.Optional(CONF_CACHE, default=DEFAULT_CACHE): cv.boolean,
|
||||
@@ -63,6 +66,7 @@ SCHEMA_SERVICE_SAY = vol.Schema({
|
||||
vol.Required(ATTR_MESSAGE): cv.string,
|
||||
vol.Optional(ATTR_ENTITY_ID): cv.entity_ids,
|
||||
vol.Optional(ATTR_CACHE): cv.boolean,
|
||||
vol.Optional(ATTR_LANGUAGE): cv.string
|
||||
})
|
||||
|
||||
SCHEMA_SERVICE_CLEAR_CACHE = vol.Schema({})
|
||||
@@ -121,10 +125,11 @@ def async_setup(hass, config):
|
||||
entity_ids = service.data.get(ATTR_ENTITY_ID)
|
||||
message = service.data.get(ATTR_MESSAGE)
|
||||
cache = service.data.get(ATTR_CACHE)
|
||||
language = service.data.get(ATTR_LANGUAGE)
|
||||
|
||||
try:
|
||||
url = yield from tts.async_get_url(
|
||||
p_type, message, cache=cache)
|
||||
p_type, message, cache=cache, language=language)
|
||||
except HomeAssistantError as err:
|
||||
_LOGGER.error("Error on init tts: %s", err)
|
||||
return
|
||||
@@ -207,7 +212,8 @@ class SpeechManager(object):
|
||||
for file_data in folder_data:
|
||||
record = _RE_VOICE_FILE.match(file_data)
|
||||
if record:
|
||||
key = "{}_{}".format(record.group(1), record.group(2))
|
||||
key = KEY_PATTERN.format(
|
||||
record.group(1), record.group(2), record.group(3))
|
||||
cache[key.lower()] = file_data.lower()
|
||||
return cache
|
||||
|
||||
@@ -241,17 +247,19 @@ class SpeechManager(object):
|
||||
def async_register_engine(self, engine, provider, config):
|
||||
"""Register a TTS provider."""
|
||||
provider.hass = self.hass
|
||||
provider.language = config.get(CONF_LANG)
|
||||
if CONF_LANG in config:
|
||||
provider.language = config.get(CONF_LANG)
|
||||
self.providers[engine] = provider
|
||||
|
||||
@asyncio.coroutine
|
||||
def async_get_url(self, engine, message, cache=None):
|
||||
def async_get_url(self, engine, message, cache=None, language=None):
|
||||
"""Get URL for play message.
|
||||
|
||||
This method is a coroutine.
|
||||
"""
|
||||
msg_hash = hashlib.sha1(bytes(message, 'utf-8')).hexdigest()
|
||||
key = ("{}_{}".format(msg_hash, engine)).lower()
|
||||
language_key = language or self.providers[engine].language
|
||||
key = KEY_PATTERN.format(msg_hash, language_key, engine).lower()
|
||||
use_cache = cache if cache is not None else self.use_cache
|
||||
|
||||
# is speech allready in memory
|
||||
@@ -260,23 +268,24 @@ class SpeechManager(object):
|
||||
# is file store in file cache
|
||||
elif use_cache and key in self.file_cache:
|
||||
filename = self.file_cache[key]
|
||||
self.hass.async_add_job(self.async_file_to_mem(engine, key))
|
||||
self.hass.async_add_job(self.async_file_to_mem(key))
|
||||
# load speech from provider into memory
|
||||
else:
|
||||
filename = yield from self.async_get_tts_audio(
|
||||
engine, key, message, use_cache)
|
||||
engine, key, message, use_cache, language)
|
||||
|
||||
return "{}/api/tts_proxy/{}".format(
|
||||
self.hass.config.api.base_url, filename)
|
||||
|
||||
@asyncio.coroutine
|
||||
def async_get_tts_audio(self, engine, key, message, cache):
|
||||
def async_get_tts_audio(self, engine, key, message, cache, language):
|
||||
"""Receive TTS and store for view in cache.
|
||||
|
||||
This method is a coroutine.
|
||||
"""
|
||||
provider = self.providers[engine]
|
||||
extension, data = yield from provider.async_get_tts_audio(message)
|
||||
extension, data = yield from provider.async_get_tts_audio(
|
||||
message, language)
|
||||
|
||||
if data is None or extension is None:
|
||||
raise HomeAssistantError(
|
||||
@@ -314,7 +323,7 @@ class SpeechManager(object):
|
||||
_LOGGER.error("Can't write %s", filename)
|
||||
|
||||
@asyncio.coroutine
|
||||
def async_file_to_mem(self, engine, key):
|
||||
def async_file_to_mem(self, key):
|
||||
"""Load voice from file cache into memory.
|
||||
|
||||
This method is a coroutine.
|
||||
@@ -362,13 +371,13 @@ class SpeechManager(object):
|
||||
if not record:
|
||||
raise HomeAssistantError("Wrong tts file format!")
|
||||
|
||||
key = "{}_{}".format(record.group(1), record.group(2))
|
||||
key = KEY_PATTERN.format(
|
||||
record.group(1), record.group(2), record.group(3))
|
||||
|
||||
if key not in self.mem_cache:
|
||||
if key not in self.file_cache:
|
||||
raise HomeAssistantError("%s not in cache!", key)
|
||||
engine = record.group(2)
|
||||
yield from self.async_file_to_mem(engine, key)
|
||||
yield from self.async_file_to_mem(key)
|
||||
|
||||
content, _ = mimetypes.guess_type(filename)
|
||||
return (content, self.mem_cache[key][MEM_CACHE_VOICE])
|
||||
@@ -380,11 +389,11 @@ class Provider(object):
|
||||
hass = None
|
||||
language = None
|
||||
|
||||
def get_tts_audio(self, message):
|
||||
def get_tts_audio(self, message, language=None):
|
||||
"""Load tts audio file from provider."""
|
||||
raise NotImplementedError()
|
||||
|
||||
def async_get_tts_audio(self, message):
|
||||
def async_get_tts_audio(self, message, language=None):
|
||||
"""Load tts audio file from provider.
|
||||
|
||||
Return a tuple of file extension and data as bytes.
|
||||
@@ -392,7 +401,8 @@ class Provider(object):
|
||||
This method must be run in the event loop and returns a coroutine.
|
||||
"""
|
||||
return self.hass.loop.run_in_executor(
|
||||
None, self.get_tts_audio, message)
|
||||
None,
|
||||
functools.partial(self.get_tts_audio, message, language=language))
|
||||
|
||||
|
||||
class TextToSpeechView(HomeAssistantView):
|
||||
|
||||
@@ -17,7 +17,11 @@ def get_engine(hass, config):
|
||||
class DemoProvider(Provider):
|
||||
"""Demo speech api provider."""
|
||||
|
||||
def get_tts_audio(self, message):
|
||||
def __init__(self):
|
||||
"""Initialize demo provider for TTS."""
|
||||
self.language = 'en'
|
||||
|
||||
def get_tts_audio(self, message, language=None):
|
||||
"""Load TTS from demo."""
|
||||
filename = os.path.join(os.path.dirname(__file__), "demo.mp3")
|
||||
try:
|
||||
|
||||
@@ -59,7 +59,7 @@ class GoogleProvider(Provider):
|
||||
}
|
||||
|
||||
@asyncio.coroutine
|
||||
def async_get_tts_audio(self, message):
|
||||
def async_get_tts_audio(self, message, language=None):
|
||||
"""Load TTS from google."""
|
||||
from gtts_token import gtts_token
|
||||
|
||||
@@ -67,6 +67,11 @@ class GoogleProvider(Provider):
|
||||
websession = async_get_clientsession(self.hass)
|
||||
message_parts = self._split_message_to_parts(message)
|
||||
|
||||
# If language is not specified or is not supported - use the language
|
||||
# from the config.
|
||||
if language not in SUPPORT_LANGUAGES:
|
||||
language = self.language
|
||||
|
||||
data = b''
|
||||
for idx, part in enumerate(message_parts):
|
||||
part_token = yield from self.hass.loop.run_in_executor(
|
||||
@@ -74,7 +79,7 @@ class GoogleProvider(Provider):
|
||||
|
||||
url_param = {
|
||||
'ie': 'UTF-8',
|
||||
'tl': self.language,
|
||||
'tl': language,
|
||||
'q': yarl.quote(part),
|
||||
'tk': part_token,
|
||||
'total': len(message_parts),
|
||||
|
||||
@@ -14,5 +14,9 @@ say:
|
||||
description: Control file cache of this message.
|
||||
example: 'true'
|
||||
|
||||
language:
|
||||
description: Language to use for speech generation.
|
||||
example: 'ru'
|
||||
|
||||
clear_cache:
|
||||
description: Remove cache files and RAM cache.
|
||||
@@ -103,13 +103,18 @@ class VoiceRSSProvider(Provider):
|
||||
}
|
||||
|
||||
@asyncio.coroutine
|
||||
def async_get_tts_audio(self, message):
|
||||
def async_get_tts_audio(self, message, language=None):
|
||||
"""Load TTS from voicerss."""
|
||||
websession = async_get_clientsession(self.hass)
|
||||
form_data = self.form_data.copy()
|
||||
|
||||
form_data['src'] = message
|
||||
|
||||
# If language is specified and supported - use it instead of the
|
||||
# language in the config.
|
||||
if language in SUPPORT_LANGUAGES:
|
||||
form_data['hl'] = language
|
||||
|
||||
request = None
|
||||
try:
|
||||
with async_timeout.timeout(10, loop=self.hass.loop):
|
||||
|
||||
Reference in new issue
Block a user