livekit-plugins-google 1.0.13__py3-none-any.whl → 1.0.15__py3-none-any.whl

livekit/plugins/google/beta/realtime/api_proto.py

@@ -5,7 +5,7 @@ from typing import Literal, Union
 
 from google.genai import types
 
-LiveAPIModels = Literal["gemini-2.0-flash-exp"]
+LiveAPIModels = Literal["gemini-2.0-flash-exp", "gemini-2.0-flash-live-001"]
 
 Voice = Literal["Puck", "Charon", "Kore", "Fenrir", "Aoede"]
 

livekit/plugins/google/beta/realtime/realtime_api.py

@@ -95,7 +95,7 @@ class RealtimeModel(llm.RealtimeModel):
         self,
         *,
         instructions: NotGivenOr[str] = NOT_GIVEN,
-        model: LiveAPIModels | str = "gemini-2.0-flash-exp",
+        model: LiveAPIModels | str = "gemini-2.0-flash-live-001",
         api_key: NotGivenOr[str] = NOT_GIVEN,
         voice: Voice | str = "Puck",
         modalities: NotGivenOr[list[Modality]] = NOT_GIVEN,
@@ -357,10 +357,10 @@ class RealtimeSession(llm.RealtimeSession):
         return fut
 
     def interrupt(self) -> None:
-        logger.warning("interrupt() - no direct cancellation in Gemini")
+        pass
 
     def truncate(self, *, message_id: str, audio_end_ms: int) -> None:
-        logger.warning(f"truncate(...) called for {message_id}, ignoring for Gemini")
+        pass
 
     async def aclose(self) -> None:
         self._msg_ch.close()
@@ -423,8 +423,8 @@ class RealtimeSession(llm.RealtimeSession):
                     async for msg in self._msg_ch:
                         if isinstance(msg, LiveClientContent):
                             await session.send(input=msg, end_of_turn=True)
-
-                        await session.send(input=msg)
+                        else:
+                            await session.send(input=msg)
                     await session.send(input=".", end_of_turn=True)
 
                 @utils.log_exceptions(logger=logger)
@@ -543,8 +543,11 @@ class RealtimeSession(llm.RealtimeSession):
         output_transcription = server_content.output_transcription
         if output_transcription and output_transcription.text:
             item_generation.text_ch.send_nowait(output_transcription.text)
+        if server_content.interrupted:
+            self._finalize_response()
+            self._handle_input_speech_started()
 
-        if server_content.interrupted or server_content.turn_complete:
+        if server_content.turn_complete:
             self._finalize_response()
 
     def _finalize_response(self) -> None:
@@ -560,7 +563,9 @@ class RealtimeSession(llm.RealtimeSession):
         self._current_generation = None
         self._is_interrupted = True
         self._active_response_id = None
-        self.emit("agent_speech_stopped")
+
+    def _handle_input_speech_started(self):
+        self.emit("input_speech_started", llm.InputSpeechStartedEvent())
 
     def _handle_tool_calls(self, tool_call: LiveServerToolCall):
         if not self._current_generation:

livekit/plugins/google/llm.py

@@ -53,6 +53,7 @@ class _LLMOptions:
     top_k: NotGivenOr[float]
     presence_penalty: NotGivenOr[float]
     frequency_penalty: NotGivenOr[float]
+    thinking_config: NotGivenOr[types.ThinkingConfigOrDict]
 
 
 class LLM(llm.LLM):
@@ -71,6 +72,7 @@ class LLM(llm.LLM):
         presence_penalty: NotGivenOr[float] = NOT_GIVEN,
         frequency_penalty: NotGivenOr[float] = NOT_GIVEN,
         tool_choice: NotGivenOr[ToolChoice] = NOT_GIVEN,
+        thinking_config: NotGivenOr[types.ThinkingConfigOrDict] = NOT_GIVEN,
     ) -> None:
         """
         Create a new instance of Google GenAI LLM.
@@ -95,6 +97,7 @@ class LLM(llm.LLM):
             presence_penalty (float, optional): Penalizes the model for generating previously mentioned concepts. Defaults to None.
             frequency_penalty (float, optional): Penalizes the model for repeating words. Defaults to None.
             tool_choice (ToolChoice, optional): Specifies whether to use tools during response generation. Defaults to "auto".
+            thinking_config (ThinkingConfigOrDict, optional): The thinking configuration for response generation. Defaults to None.
         """  # noqa: E501
         super().__init__()
         gcp_project = project if is_given(project) else os.environ.get("GOOGLE_CLOUD_PROJECT")
@@ -121,6 +124,22 @@ class LLM(llm.LLM):
                     "API key is required for Google API either via api_key or GOOGLE_API_KEY environment variable"  # noqa: E501
                 )
 
+        # Validate thinking_config
+        if is_given(thinking_config):
+            _thinking_budget = None
+            if isinstance(thinking_config, dict):
+                _thinking_budget = thinking_config.get("thinking_budget")
+            elif isinstance(thinking_config, types.ThinkingConfig):
+                _thinking_budget = thinking_config.thinking_budget
+
+            if _thinking_budget is not None:
+                if not isinstance(_thinking_budget, int):
+                    raise ValueError("thinking_budget inside thinking_config must be an integer")
+                if not (0 <= _thinking_budget <= 24576):
+                    raise ValueError(
+                        "thinking_budget inside thinking_config must be between 0 and 24576"
+                    )
+
         self._opts = _LLMOptions(
             model=model,
             temperature=temperature,
@@ -133,6 +152,7 @@ class LLM(llm.LLM):
             top_k=top_k,
             presence_penalty=presence_penalty,
             frequency_penalty=frequency_penalty,
+            thinking_config=thinking_config,
         )
         self._client = genai.Client(
             api_key=gemini_api_key,
@@ -212,6 +232,10 @@ class LLM(llm.LLM):
         if is_given(self._opts.frequency_penalty):
             extra["frequency_penalty"] = self._opts.frequency_penalty
 
+        # Add thinking config if thinking_budget is provided
+        if is_given(self._opts.thinking_config):
+            extra["thinking_config"] = self._opts.thinking_config
+
         return LLMStream(
             self,
             client=self._client,
@@ -310,7 +334,7 @@ class LLMStream(llm.LLMStream):
             raise APIStatusError(
                 "gemini llm: client error",
                 status_code=e.code,
-                body=e.message + e.status,
+                body=f"{e.message} {e.status}",
                 request_id=request_id,
                 retryable=False if e.code != 429 else True,
             ) from e
@@ -318,7 +342,7 @@ class LLMStream(llm.LLMStream):
             raise APIStatusError(
                 "gemini llm: server error",
                 status_code=e.code,
-                body=e.message + e.status,
+                body=f"{e.message} {e.status}",
                 request_id=request_id,
                 retryable=retryable,
             ) from e
@@ -326,7 +350,7 @@ class LLMStream(llm.LLMStream):
             raise APIStatusError(
                 "gemini llm: api error",
                 status_code=e.code,
-                body=e.message + e.status,
+                body=f"{e.message} {e.status}",
                 request_id=request_id,
                 retryable=retryable,
             ) from e

livekit/plugins/google/tts.py

@@ -35,6 +35,8 @@ from livekit.agents.types import (
 )
 from livekit.agents.utils import is_given
 
+from .models import Gender, SpeechLanguages
+
 
 @dataclass
 class _TTSOptions:
@@ -46,7 +48,9 @@ class TTS(tts.TTS):
     def __init__(
         self,
         *,
-        voice: NotGivenOr[texttospeech.VoiceSelectionParams] = NOT_GIVEN,
+        language: NotGivenOr[SpeechLanguages | str] = NOT_GIVEN,
+        gender: NotGivenOr[Gender | str] = NOT_GIVEN,
+        voice_name: NotGivenOr[str] = NOT_GIVEN,
         sample_rate: int = 24000,
         pitch: int = 0,
         effects_profile_id: str = "",
@@ -63,7 +67,9 @@ class TTS(tts.TTS):
         environmental variable.
 
         Args:
-            voice (texttospeech.VoiceSelectionParams, optional): Voice selection parameters.
+            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.
             sample_rate (int, optional): Audio sample rate in Hz. Default is 24000.
             location (str, optional): Location for the TTS client. Default is "global".
             pitch (float, optional): Speaking pitch, ranging from -20.0 to 20.0 semitones relative to the original pitch. Default is 0.
@@ -85,15 +91,19 @@ class TTS(tts.TTS):
         self._credentials_info = credentials_info
         self._credentials_file = credentials_file
         self._location = location
-        if not is_given(voice):
-            voice = texttospeech.VoiceSelectionParams(
-                name="",
-                language_code="en-US",
-                ssml_gender=SsmlVoiceGender.NEUTRAL,
-            )
+
+        lang = language if is_given(language) else "en-US"
+        ssml_gender = _gender_from_str("neutral" if not is_given(gender) else gender)
+        name = "" if not is_given(voice_name) else voice_name
+
+        voice_params = texttospeech.VoiceSelectionParams(
+            name=name,
+            language_code=lang,
+            ssml_gender=ssml_gender,
+        )
 
         self._opts = _TTSOptions(
-            voice=voice,
+            voice=voice_params,
             audio_config=texttospeech.AudioConfig(
                 audio_encoding=texttospeech.AudioEncoding.OGG_OPUS,
                 sample_rate_hertz=sample_rate,
@@ -106,18 +116,31 @@ class TTS(tts.TTS):
     def update_options(
         self,
         *,
-        voice: NotGivenOr[texttospeech.VoiceSelectionParams] = NOT_GIVEN,
+        language: NotGivenOr[SpeechLanguages | str] = NOT_GIVEN,
+        gender: NotGivenOr[Gender | str] = NOT_GIVEN,
+        voice_name: NotGivenOr[str] = NOT_GIVEN,
         speaking_rate: NotGivenOr[float] = NOT_GIVEN,
     ) -> None:
         """
         Update the TTS options.
 
         Args:
-            voice (texttospeech.VoiceSelectionParams, optional): Voice selection parameters.
+            language (SpeechLanguages | str, optional): Language code (e.g., "en-US").
+            gender (Gender | str, optional): Voice gender ("male", "female", "neutral").
+            voice_name (str, optional): Specific voice name.
             speaking_rate (float, optional): Speed of speech.
         """  # noqa: E501
-        if is_given(voice):
-            self._opts.voice = voice
+        params = {}
+        if is_given(language):
+            params["language"] = language
+        if is_given(gender):
+            params["gender"] = gender
+        if is_given(voice_name):
+            params["voice_name"] = voice_name
+
+        if params:
+            self._opts.voice = texttospeech.VoiceSelectionParams(**params)
+
         if is_given(speaking_rate):
             self._opts.audio_config.speaking_rate = speaking_rate
 

livekit/plugins/google/version.py

@@ -12,4 +12,4 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-__version__ = "1.0.13"
+__version__ = "1.0.15"

{livekit_plugins_google-1.0.13.dist-info → livekit_plugins_google-1.0.15.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: livekit-plugins-google
-Version: 1.0.13
+Version: 1.0.15
 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
 Requires-Dist: google-genai>=1.10.0
-Requires-Dist: livekit-agents>=1.0.13
+Requires-Dist: livekit-agents>=1.0.15
 Description-Content-Type: text/markdown
 
 # LiveKit Plugins Google

{livekit_plugins_google-1.0.13.dist-info → livekit_plugins_google-1.0.15.dist-info}/RECORD

@@ -1,16 +1,16 @@
 livekit/plugins/google/__init__.py,sha256=e_kSlFNmKhyyeliz7f4WOKc_Y0-y39QjO5nCWuguhss,1171
-livekit/plugins/google/llm.py,sha256=yAm-to2ItTJ7dAHc-2mlPeI0Npz9ZxRdyuRLV8PINqg,14888
+livekit/plugins/google/llm.py,sha256=SqNGg6-wlrIUo9uaismP7QW5XztkXyDivJXLVgOIZMI,16175
 livekit/plugins/google/log.py,sha256=GI3YWN5YzrafnUccljzPRS_ZALkMNk1i21IRnTl2vNA,69
 livekit/plugins/google/models.py,sha256=SGjAumdDK97NNLwMFcqZdKR68f1NoGB2Rk1UP2-imG0,1457
 livekit/plugins/google/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 livekit/plugins/google/stt.py,sha256=AG_lh2fuuduJi0jFbA_QKFXLJ6NUdF1W_FfkLUJML_Q,22413
-livekit/plugins/google/tts.py,sha256=P8Zu2s0TfmyzlrNxzDIqyn3sGiNSW0n3nB_JlO_ojiM,7985
+livekit/plugins/google/tts.py,sha256=xhINokqY8UutXn85N-cbzq68eptbM6TTtIXmLktE_RM,9004
 livekit/plugins/google/utils.py,sha256=pbLSOAdQxInWhgI2Yhsrr9KvgvpFXYDdU2yx2p03pFg,9437
-livekit/plugins/google/version.py,sha256=i9Tq4ZlIN5uba7xHRxp31dxAE9NuzqobM8zWhdM4QgA,601
+livekit/plugins/google/version.py,sha256=wHPUkZRYx-OB6iDuwTmMNVVQXU9eg5xFSjgmKBqqwd4,601
 livekit/plugins/google/beta/__init__.py,sha256=AxRYc7NGG62Tv1MmcZVCDHNvlhbC86hM-_yP01Qb28k,47
 livekit/plugins/google/beta/realtime/__init__.py,sha256=_fW2NMN22F-hnQ4xAJ_g5lPbR7CvM_xXzSWlUQY-E-U,188
-livekit/plugins/google/beta/realtime/api_proto.py,sha256=cwpFOYjN_3v5PMY0TnzoHhJoASfZ7Qt9IO281ZhJ7Ww,565
-livekit/plugins/google/beta/realtime/realtime_api.py,sha256=JBEEOeTl6gv6Fe6GtYJjj9C-dqvfhWpOzNAa0tnTKgM,25002
-livekit_plugins_google-1.0.13.dist-info/METADATA,sha256=u8ocRjsu24AzO_FRgqYZzDqc3gKnQGp1hprKBc3RFm4,3492
-livekit_plugins_google-1.0.13.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-livekit_plugins_google-1.0.13.dist-info/RECORD,,
+livekit/plugins/google/beta/realtime/api_proto.py,sha256=Fyrejs3SG0EjOPCCFLEnWXKEUxCff47PMWk2VsKJm5E,594
+livekit/plugins/google/beta/realtime/realtime_api.py,sha256=HvPYyQXC9OodWaDNxbRt1UAJ8IVdXZGK-PsIEr7UwbY,25078
+livekit_plugins_google-1.0.15.dist-info/METADATA,sha256=wMOLBkgHx_fJ0o5s8URB7Ev6yEg2jhKHhb0OlH1_7p4,3492
+livekit_plugins_google-1.0.15.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+livekit_plugins_google-1.0.15.dist-info/RECORD,,