livekit-plugins-google 1.1.7__tar.gz → 1.2.1__tar.gz

{livekit_plugins_google-1.1.7 → livekit_plugins_google-1.2.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: livekit-plugins-google
-Version: 1.1.7
+Version: 1.2.1
 Summary: Agent Framework plugin for services from Google Cloud
 Project-URL: Documentation, https://docs.livekit.io
 Project-URL: Website, https://livekit.io/
@@ -22,7 +22,7 @@ Requires-Dist: google-auth<3,>=2
 Requires-Dist: google-cloud-speech<3,>=2
 Requires-Dist: google-cloud-texttospeech<3,>=2.27
 Requires-Dist: google-genai>=v1.23.0
-Requires-Dist: livekit-agents>=1.1.7
+Requires-Dist: livekit-agents>=1.2.1
 Description-Content-Type: text/markdown
 
 # Google AI plugin for LiveKit Agents

{livekit_plugins_google-1.1.7 → livekit_plugins_google-1.2.1}/livekit/plugins/google/beta/realtime/realtime_api.py

@@ -52,7 +52,7 @@ class _RealtimeOptions:
     api_key: str | None
     voice: Voice | str
     language: NotGivenOr[str]
-    response_modalities: NotGivenOr[list[types.Modality]]
+    response_modalities: list[types.Modality]
     vertexai: bool
     project: str | None
     location: str | None
@@ -192,12 +192,15 @@ class RealtimeModel(llm.RealtimeModel):
         ):
             server_turn_detection = False
 
+        modalities = modalities if is_given(modalities) else [types.Modality.AUDIO]
+
         super().__init__(
             capabilities=llm.RealtimeCapabilities(
                 message_truncation=False,
                 turn_detection=server_turn_detection,
                 user_transcription=input_audio_transcription is not None,
                 auto_tool_reply_generation=True,
+                audio_output=types.Modality.AUDIO in modalities,
             )
         )
 
@@ -538,7 +541,9 @@ class RealtimeSession(llm.RealtimeSession):
             return
         self.start_user_activity()
 
-    def truncate(self, *, message_id: str, audio_end_ms: int) -> None:
+    def truncate(
+        self, *, message_id: str, audio_end_ms: int, audio_transcript: NotGivenOr[str] = NOT_GIVEN
+    ) -> None:
         logger.warning("truncate is not supported by the Google Realtime API.")
         pass
 
@@ -732,9 +737,7 @@ class RealtimeSession(llm.RealtimeSession):
             gemini_tools=self._opts.gemini_tools if is_given(self._opts.gemini_tools) else None,
         )
         conf = types.LiveConnectConfig(
-            response_modalities=self._opts.response_modalities
-            if is_given(self._opts.response_modalities)
-            else [types.Modality.AUDIO],
+            response_modalities=self._opts.response_modalities,
             generation_config=types.GenerationConfig(
                 candidate_count=self._opts.candidate_count,
                 temperature=temp,
@@ -793,6 +796,8 @@ class RealtimeSession(llm.RealtimeSession):
             audio_ch=utils.aio.Chan[rtc.AudioFrame](),
             _created_timestamp=time.time(),
         )
+        if not self._realtime_model.capabilities.audio_output:
+            self._current_generation.audio_ch.close()
 
         self._current_generation.message_ch.send_nowait(
             llm.MessageGeneration(

{livekit_plugins_google-1.1.7 → livekit_plugins_google-1.2.1}/livekit/plugins/google/llm.py

@@ -187,6 +187,10 @@ class LLM(llm.LLM):
             location=gcp_location,
         )
 
+    @property
+    def model(self) -> str:
+        return self._opts.model
+
     def chat(
         self,
         *,

{livekit_plugins_google-1.1.7 → livekit_plugins_google-1.2.1}/livekit/plugins/google/tts.py

@@ -34,7 +34,6 @@ from livekit.agents.utils import is_given
 from .log import logger
 from .models import Gender, SpeechLanguages
 
-BUFFERED_WORDS_COUNT = 8
 NUM_CHANNELS = 1
 DEFAULT_VOICE_NAME = "en-US-Chirp3-HD-Charon"
 DEFAULT_LANGUAGE = "en-US"
@@ -124,7 +123,7 @@ class TTS(tts.TTS):
             ssml_gender=ssml_gender,
         )
         if not is_given(tokenizer):
-            tokenizer = tokenize.basic.SentenceTokenizer(min_sentence_len=BUFFERED_WORDS_COUNT)
+            tokenizer = tokenize.blingfire.SentenceTokenizer()
 
         pronunciations = None if not is_given(custom_pronunciations) else custom_pronunciations
 

{livekit_plugins_google-1.1.7 → livekit_plugins_google-1.2.1}/livekit/plugins/google/version.py

@@ -12,4 +12,4 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-__version__ = "1.1.7"
+__version__ = "1.2.1"

{livekit_plugins_google-1.1.7 → livekit_plugins_google-1.2.1}/pyproject.toml

@@ -27,7 +27,7 @@ dependencies = [
     "google-cloud-speech >= 2, < 3",
     "google-cloud-texttospeech >= 2.27, < 3",
     "google-genai >= v1.23.0",
-    "livekit-agents>=1.1.7",
+    "livekit-agents>=1.2.1",
 ]
 
 [project.urls]