livekit-plugins-google 0.7.1__py3-none-any.whl → 0.7.3__py3-none-any.whl

livekit/plugins/google/__init__.py

@@ -29,3 +29,12 @@ class GooglePlugin(Plugin):
 
 
 Plugin.register_plugin(GooglePlugin())
+
+# Cleanup docs of unexported modules
+_module = dir()
+NOT_IN_ALL = [m for m in _module if m not in __all__]
+
+__pdoc__ = {}
+
+for n in NOT_IN_ALL:
+    __pdoc__[n] = False

livekit/plugins/google/stt.py

@@ -20,8 +20,15 @@ from dataclasses import dataclass
 from typing import AsyncIterable, List, Union
 
 from livekit import agents, rtc
-from livekit.agents import stt, utils
-
+from livekit.agents import (
+    APIConnectionError,
+    APIStatusError,
+    APITimeoutError,
+    stt,
+    utils,
+)
+
+from google.api_core.exceptions import Aborted, DeadlineExceeded, GoogleAPICallError
 from google.auth import default as gauth_default
 from google.auth.exceptions import DefaultCredentialsError
 from google.cloud.speech_v2 import SpeechAsyncClient
@@ -43,6 +50,25 @@ class STTOptions:
     punctuate: bool
     spoken_punctuation: bool
     model: SpeechModels
+    keywords: List[tuple[str, float]] | None
+
+    def build_adaptation(self) -> cloud_speech.SpeechAdaptation | None:
+        if self.keywords:
+            return cloud_speech.SpeechAdaptation(
+                phrase_sets=[
+                    cloud_speech.SpeechAdaptation.AdaptationPhraseSet(
+                        inline_phrase_set=cloud_speech.PhraseSet(
+                            phrases=[
+                                cloud_speech.PhraseSet.Phrase(
+                                    value=keyword, boost=boost
+                                )
+                                for keyword, boost in self.keywords
+                            ]
+                        )
+                    )
+                ]
+            )
+        return None
 
 
 class STT(stt.STT):
@@ -57,6 +83,7 @@ class STT(stt.STT):
         model: SpeechModels = "long",
         credentials_info: dict | None = None,
         credentials_file: str | None = None,
+        keywords: List[tuple[str, float]] | None = None,
     ):
         """
         Create a new instance of Google STT.
@@ -93,6 +120,7 @@ class STT(stt.STT):
             punctuate=punctuate,
             spoken_punctuation=spoken_punctuation,
             model=model,
+            keywords=keywords,
         )
 
     def _ensure_client(self) -> SpeechAsyncClient:
@@ -141,7 +169,7 @@ class STT(stt.STT):
 
         return config
 
-    async def recognize(
+    async def _recognize_impl(
         self,
         buffer: utils.AudioBuffer,
         *,
@@ -156,6 +184,7 @@ class STT(stt.STT):
                 sample_rate_hertz=frame.sample_rate,
                 audio_channel_count=frame.num_channels,
             ),
+            adaptation=config.build_adaptation(),
             features=cloud_speech.RecognitionFeatures(
                 enable_automatic_punctuation=config.punctuate,
                 enable_spoken_punctuation=config.spoken_punctuation,
@@ -165,23 +194,39 @@ class STT(stt.STT):
             language_codes=config.languages,
         )
 
-        raw = await self._ensure_client().recognize(
-            cloud_speech.RecognizeRequest(
-                recognizer=self._recognizer, config=config, content=frame.data.tobytes()
+        try:
+            raw = await self._ensure_client().recognize(
+                cloud_speech.RecognizeRequest(
+                    recognizer=self._recognizer,
+                    config=config,
+                    content=frame.data.tobytes(),
+                )
             )
-        )
-        return _recognize_response_to_speech_event(raw)
+
+            return _recognize_response_to_speech_event(raw)
+        except DeadlineExceeded:
+            raise APITimeoutError()
+        except GoogleAPICallError as e:
+            raise APIStatusError(
+                e.message,
+                status_code=e.code or -1,
+                request_id=None,
+                body=None,
+            )
+        except Exception as e:
+            raise APIConnectionError() from e
 
     def stream(
         self, *, language: SpeechLanguages | str | None = None
     ) -> "SpeechStream":
         config = self._sanitize_options(language=language)
-        return SpeechStream(self._ensure_client(), self._recognizer, config)
+        return SpeechStream(self, self._ensure_client(), self._recognizer, config)
 
 
 class SpeechStream(stt.SpeechStream):
     def __init__(
         self,
+        stt: STT,
         client: SpeechAsyncClient,
         recognizer: str,
         config: STTOptions,
@@ -189,7 +234,7 @@ class SpeechStream(stt.SpeechStream):
         num_channels: int = 1,
         max_retry: int = 32,
     ) -> None:
-        super().__init__()
+        super().__init__(stt)
 
         self._client = client
         self._recognizer = recognizer
@@ -205,6 +250,7 @@ class SpeechStream(stt.SpeechStream):
                     sample_rate_hertz=self._sample_rate,
                     audio_channel_count=self._num_channels,
                 ),
+                adaptation=config.build_adaptation(),
                 language_codes=self._config.languages,
                 model=self._config.model,
                 features=cloud_speech.RecognitionFeatures(
@@ -257,6 +303,9 @@ class SpeechStream(stt.SpeechStream):
                 retry_count = 0  # connection successful, reset retry count
 
                 await self._run_stream(stream)
+            except Aborted:
+                logger.error("google stt connection aborted")
+                break
             except Exception as e:
                 if retry_count >= max_retry:
                     logger.error(

livekit/plugins/google/tts.py

@@ -15,21 +15,22 @@
 from __future__ import annotations
 
 from dataclasses import dataclass
-from typing import Union
 
 from livekit import rtc
-from livekit.agents import tts, utils
-
+from livekit.agents import (
+    APIConnectionError,
+    APIStatusError,
+    APITimeoutError,
+    tts,
+    utils,
+)
+
+from google.api_core.exceptions import DeadlineExceeded, GoogleAPICallError
 from google.cloud import texttospeech
 from google.cloud.texttospeech_v1.types import SsmlVoiceGender, SynthesizeSpeechResponse
 
-from .log import logger
 from .models import AudioEncoding, Gender, SpeechLanguages
 
-LgType = Union[SpeechLanguages, str]
-GenderType = Union[Gender, str]
-AudioEncodingType = Union[AudioEncoding, str]
-
 
 @dataclass
 class _TTSOptions:
@@ -41,11 +42,13 @@ class TTS(tts.TTS):
     def __init__(
         self,
         *,
-        language: LgType = "en-US",
-        gender: GenderType = "neutral",
+        language: SpeechLanguages | str = "en-US",
+        gender: Gender | str = "neutral",
         voice_name: str = "",  # Not required
-        encoding: AudioEncodingType = "linear16",
+        encoding: AudioEncoding | str = "linear16",
         sample_rate: int = 24000,
+        pitch: int = 0,
+        effects_profile_id: str = "",
         speaking_rate: float = 1.0,
         credentials_info: dict | None = None,
         credentials_file: str | None = None,
@@ -56,6 +59,18 @@ class TTS(tts.TTS):
         Credentials must be provided, either by using the ``credentials_info`` dict, or reading
         from the file specified in ``credentials_file`` or the ``GOOGLE_APPLICATION_CREDENTIALS``
         environmental variable.
+
+        Args:
+            language (SpeechLanguages | str, optional): Language code (e.g., "en-US"). Default is "en-US".
+            gender (Gender | str, optional): Voice gender ("male", "female", "neutral"). Default is "neutral".
+            voice_name (str, optional): Specific voice name. Default is an empty string.
+            encoding (AudioEncoding | str, optional): Audio encoding format (e.g., "linear16"). Default is "linear16".
+            sample_rate (int, optional): Audio sample rate in Hz. Default is 24000.
+            pitch (float, optional): Speaking pitch, ranging from -20.0 to 20.0 semitones relative to the original pitch. Default is 0.
+            effects_profile_id (str): Optional identifier for selecting audio effects profiles to apply to the synthesized speech.
+            speaking_rate (float, optional): Speed of speech. Default is 1.0.
+            credentials_info (dict, optional): Dictionary containing Google Cloud credentials. Default is None.
+            credentials_file (str, optional): Path to the Google Cloud credentials JSON file. Default is None.
         """
 
         super().__init__(
@@ -70,14 +85,10 @@ class TTS(tts.TTS):
         self._credentials_info = credentials_info
         self._credentials_file = credentials_file
 
-        ssml_gender = SsmlVoiceGender.NEUTRAL
-        if gender == "male":
-            ssml_gender = SsmlVoiceGender.MALE
-        elif gender == "female":
-            ssml_gender = SsmlVoiceGender.FEMALE
-
         voice = texttospeech.VoiceSelectionParams(
-            name=voice_name, language_code=language, ssml_gender=ssml_gender
+            name=voice_name,
+            language_code=language,
+            ssml_gender=_gender_from_str(gender),
         )
 
         if encoding == "linear16" or encoding == "wav":
@@ -92,10 +103,36 @@ class TTS(tts.TTS):
             audio_config=texttospeech.AudioConfig(
                 audio_encoding=_audio_encoding,
                 sample_rate_hertz=sample_rate,
+                pitch=pitch,
+                effects_profile_id=effects_profile_id,
                 speaking_rate=speaking_rate,
             ),
         )
 
+    def update_options(
+        self,
+        *,
+        language: SpeechLanguages | str = "en-US",
+        gender: Gender | str = "neutral",
+        voice_name: str = "",  # Not required
+        speaking_rate: float = 1.0,
+    ) -> None:
+        """
+        Update the TTS options.
+
+        Args:
+            language (SpeechLanguages | str, optional): Language code (e.g., "en-US"). Default is "en-US".
+            gender (Gender | str, optional): Voice gender ("male", "female", "neutral"). Default is "neutral".
+            voice_name (str, optional): Specific voice name. Default is an empty string.
+            speaking_rate (float, optional): Speed of speech. Default is 1.0.
+        """
+        self._opts.voice = texttospeech.VoiceSelectionParams(
+            name=voice_name,
+            language_code=language,
+            ssml_gender=_gender_from_str(gender),
+        )
+        self._opts.audio_config.speaking_rate = speaking_rate
+
     def _ensure_client(self) -> texttospeech.TextToSpeechAsyncClient:
         if not self._client:
             if self._credentials_info:
@@ -118,57 +155,79 @@ class TTS(tts.TTS):
         return self._client
 
     def synthesize(self, text: str) -> "ChunkedStream":
-        return ChunkedStream(text, self._opts, self._ensure_client())
+        return ChunkedStream(self, text, self._opts, self._ensure_client())
 
 
 class ChunkedStream(tts.ChunkedStream):
     def __init__(
-        self, text: str, opts: _TTSOptions, client: texttospeech.TextToSpeechAsyncClient
+        self,
+        tts: TTS,
+        text: str,
+        opts: _TTSOptions,
+        client: texttospeech.TextToSpeechAsyncClient,
     ) -> None:
-        super().__init__()
-        self._text, self._opts, self._client = text, opts, client
+        super().__init__(tts, text)
+        self._opts, self._client = opts, client
 
-    @utils.log_exceptions(logger=logger)
     async def _main_task(self) -> None:
         request_id = utils.shortuuid()
-        segment_id = utils.shortuuid()
-        response: SynthesizeSpeechResponse = await self._client.synthesize_speech(
-            input=texttospeech.SynthesisInput(text=self._text),
-            voice=self._opts.voice,
-            audio_config=self._opts.audio_config,
-        )
 
-        data = response.audio_content
-        if self._opts.audio_config.audio_encoding == "mp3":
-            decoder = utils.codecs.Mp3StreamDecoder()
-            bstream = utils.audio.AudioByteStream(
-                sample_rate=self._opts.audio_config.sample_rate_hertz, num_channels=1
+        try:
+            response: SynthesizeSpeechResponse = await self._client.synthesize_speech(
+                input=texttospeech.SynthesisInput(text=self._input_text),
+                voice=self._opts.voice,
+                audio_config=self._opts.audio_config,
             )
-            for frame in decoder.decode_chunk(data):
-                for frame in bstream.write(frame.data):
-                    self._event_ch.send_nowait(
-                        tts.SynthesizedAudio(
-                            request_id=request_id, segment_id=segment_id, frame=frame
+
+            data = response.audio_content
+            if self._opts.audio_config.audio_encoding == "mp3":
+                decoder = utils.codecs.Mp3StreamDecoder()
+                bstream = utils.audio.AudioByteStream(
+                    sample_rate=self._opts.audio_config.sample_rate_hertz,
+                    num_channels=1,
+                )
+                for frame in decoder.decode_chunk(data):
+                    for frame in bstream.write(frame.data.tobytes()):
+                        self._event_ch.send_nowait(
+                            tts.SynthesizedAudio(request_id=request_id, frame=frame)
                         )
-                    )
 
-            for frame in bstream.flush():
+                for frame in bstream.flush():
+                    self._event_ch.send_nowait(
+                        tts.SynthesizedAudio(request_id=request_id, frame=frame)
+                    )
+            else:
+                data = data[44:]  # skip WAV header
                 self._event_ch.send_nowait(
                     tts.SynthesizedAudio(
-                        request_id=request_id, segment_id=segment_id, frame=frame
+                        request_id=request_id,
+                        frame=rtc.AudioFrame(
+                            data=data,
+                            sample_rate=self._opts.audio_config.sample_rate_hertz,
+                            num_channels=1,
+                            samples_per_channel=len(data) // 2,  # 16-bit
+                        ),
                     )
                 )
-        else:
-            data = data[44:]  # skip WAV header
-            self._event_ch.send_nowait(
-                tts.SynthesizedAudio(
-                    request_id=request_id,
-                    segment_id=segment_id,
-                    frame=rtc.AudioFrame(
-                        data=data,
-                        sample_rate=self._opts.audio_config.sample_rate_hertz,
-                        num_channels=1,
-                        samples_per_channel=len(data) // 2,  # 16-bit
-                    ),
-                )
+
+        except DeadlineExceeded:
+            raise APITimeoutError()
+        except GoogleAPICallError as e:
+            raise APIStatusError(
+                e.message,
+                status_code=e.code or -1,
+                request_id=None,
+                body=None,
             )
+        except Exception as e:
+            raise APIConnectionError() from e
+
+
+def _gender_from_str(gender: str) -> SsmlVoiceGender:
+    ssml_gender = SsmlVoiceGender.NEUTRAL
+    if gender == "male":
+        ssml_gender = SsmlVoiceGender.MALE
+    elif gender == "female":
+        ssml_gender = SsmlVoiceGender.FEMALE
+
+    return ssml_gender  # type: ignore

livekit/plugins/google/version.py

@@ -12,4 +12,4 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-__version__ = "0.7.1"
+__version__ = "0.7.3"

{livekit_plugins_google-0.7.1.dist-info → livekit_plugins_google-0.7.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: livekit-plugins-google
-Version: 0.7.1
+Version: 0.7.3
 Summary: Agent Framework plugin for services from Google Cloud
 Home-page: https://github.com/livekit/agents
 License: Apache-2.0
@@ -22,7 +22,7 @@ Description-Content-Type: text/markdown
 Requires-Dist: google-auth <3,>=2
 Requires-Dist: google-cloud-speech <3,>=2
 Requires-Dist: google-cloud-texttospeech <3,>=2
-Requires-Dist: livekit-agents >=0.8.0.dev0
+Requires-Dist: livekit-agents >=0.11
 
 # LiveKit Plugins Google
 

livekit_plugins_google-0.7.3.dist-info/RECORD

@@ -0,0 +1,11 @@
+livekit/plugins/google/__init__.py,sha256=rqV6C5mFNDFlrA2IcGJrsebr2VxQwMzoDUjY1JhMBZM,1117
+livekit/plugins/google/log.py,sha256=GI3YWN5YzrafnUccljzPRS_ZALkMNk1i21IRnTl2vNA,69
+livekit/plugins/google/models.py,sha256=n8pgTJ7xyJpPCZJ_y0GzaQq6LqYknL6K6trpi07-AxM,1307
+livekit/plugins/google/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+livekit/plugins/google/stt.py,sha256=WjeqYsunW8jY-WHlnNeks7gR-TiojMRR7LYdAVdCxqY,15268
+livekit/plugins/google/tts.py,sha256=hRN8ul1lDXU8LPVEfbTszgBiRYsifZXCPMwk-Pv2KeA,8793
+livekit/plugins/google/version.py,sha256=yJeG0VwiekDJAk7GHcIAe43ebagJgloe-ZsqEGZnqzE,600
+livekit_plugins_google-0.7.3.dist-info/METADATA,sha256=8UvORpoVunOTq0xKxHEk8M3sexKFnBnu66DkEJCnrRY,1647
+livekit_plugins_google-0.7.3.dist-info/WHEEL,sha256=R06PA3UVYHThwHvxuRWMqaGcr-PuniXahwjmQRFMEkY,91
+livekit_plugins_google-0.7.3.dist-info/top_level.txt,sha256=OoDok3xUmXbZRvOrfvvXB-Juu4DX79dlq188E19YHoo,8
+livekit_plugins_google-0.7.3.dist-info/RECORD,,

{livekit_plugins_google-0.7.1.dist-info → livekit_plugins_google-0.7.3.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (75.1.0)
+Generator: setuptools (75.5.0)
 Root-Is-Purelib: true
 Tag: py3-none-any
 

livekit_plugins_google-0.7.1.dist-info/RECORD

@@ -1,11 +0,0 @@
-livekit/plugins/google/__init__.py,sha256=CYbSmm5fEw71F_r_4pEApGaWQ_r15Y3ZEocH88a4yc8,948
-livekit/plugins/google/log.py,sha256=GI3YWN5YzrafnUccljzPRS_ZALkMNk1i21IRnTl2vNA,69
-livekit/plugins/google/models.py,sha256=n8pgTJ7xyJpPCZJ_y0GzaQq6LqYknL6K6trpi07-AxM,1307
-livekit/plugins/google/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-livekit/plugins/google/stt.py,sha256=XXDOISg-8U1MzVu543xLEB3-mr_NFKJp9qo1-ya2-Hc,13569
-livekit/plugins/google/tts.py,sha256=T9AHsxofwo3XaMciJPWh9O7lTZqDVYdQQlnFPiGWVbQ,6170
-livekit/plugins/google/version.py,sha256=JOBYrlKcxbTTRXkUKH0921GsmV-i71_KHczg2cgQiLc,600
-livekit_plugins_google-0.7.1.dist-info/METADATA,sha256=MyDLqZp1DC52KWx_Re3Hj0kO75l-Dg9z9IfiihtH4KY,1653
-livekit_plugins_google-0.7.1.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
-livekit_plugins_google-0.7.1.dist-info/top_level.txt,sha256=OoDok3xUmXbZRvOrfvvXB-Juu4DX79dlq188E19YHoo,8
-livekit_plugins_google-0.7.1.dist-info/RECORD,,