livekit-plugins-google 1.0.0rc9__py3-none-any.whl → 1.0.2__py3-none-any.whl

livekit/plugins/google/stt.py

@@ -132,11 +132,11 @@ class STT(stt.STT):
             try:
                 gauth_default()
             except DefaultCredentialsError:
-                raise ValueError(  # noqa: B904
+                raise ValueError(
                     "Application default credentials must be available "
                     "when using Google STT without explicitly passing "
                     "credentials through credentials_info or credentials_file."
-                )
+                ) from None
 
         if isinstance(languages, str):
             languages = [languages]
@@ -244,12 +244,9 @@ class STT(stt.STT):
 
                 return _recognize_response_to_speech_event(raw)
         except DeadlineExceeded:
-            raise APITimeoutError()  # noqa: B904
+            raise APITimeoutError() from None
         except GoogleAPICallError as e:
-            raise APIStatusError(  # noqa: B904
-                e.message,
-                status_code=e.code or -1,
-            )
+            raise APIStatusError(e.message, status_code=e.code or -1) from None
         except Exception as e:
             raise APIConnectionError() from e
 
@@ -495,12 +492,9 @@ class SpeechStream(stt.SpeechStream):
                         await utils.aio.gracefully_cancel(process_stream_task, wait_reconnect_task)
                         should_stop.set()
             except DeadlineExceeded:
-                raise APITimeoutError()  # noqa: B904
+                raise APITimeoutError() from None
             except GoogleAPICallError as e:
-                raise APIStatusError(  # noqa: B904
-                    e.message,
-                    status_code=e.code or -1,
-                )
+                raise APIStatusError(e.message, status_code=e.code or -1) from None
             except Exception as e:
                 raise APIConnectionError() from e
 

livekit/plugins/google/tts.py

@@ -203,14 +203,11 @@ class ChunkedStream(tts.ChunkedStream):
                 await decoder.aclose()
 
         except DeadlineExceeded:
-            raise APITimeoutError()  # noqa: B904
+            raise APITimeoutError() from None
         except GoogleAPICallError as e:
-            raise APIStatusError(  # noqa: B904
-                e.message,
-                status_code=e.code or -1,
-                request_id=None,
-                body=None,
-            )
+            raise APIStatusError(
+                e.message, status_code=e.code or -1, request_id=None, body=None
+            ) from None
         except Exception as e:
             raise APIConnectionError() from e
 

livekit/plugins/google/utils.py

@@ -20,8 +20,26 @@ def to_fnc_ctx(fncs: list[FunctionTool]) -> list[types.FunctionDeclaration]:
     return [_build_gemini_fnc(fnc) for fnc in fncs]
 
 
+def get_tool_results_for_realtime(chat_ctx: llm.ChatContext) -> types.LiveClientToolResponse | None:
+    function_responses: list[types.FunctionResponse] = []
+    for msg in chat_ctx.items:
+        if msg.type == "function_call_output":
+            function_responses.append(
+                types.FunctionResponse(
+                    id=msg.call_id,
+                    name=msg.name,
+                    response={"text": msg.output},
+                )
+            )
+    return (
+        types.LiveClientToolResponse(function_responses=function_responses)
+        if function_responses
+        else None
+    )
+
+
 def to_chat_ctx(
-    chat_ctx: llm.ChatContext, cache_key: Any
+    chat_ctx: llm.ChatContext, cache_key: Any, ignore_functions: bool = False
 ) -> tuple[list[types.Content], types.Content | None]:
     turns: list[types.Content] = []
     system_instruction: types.Content | None = None
@@ -59,7 +77,7 @@ def to_chat_ctx(
                     parts.append(types.Part(text=json.dumps(content)))
                 elif isinstance(content, llm.ImageContent):
                     parts.append(_to_image_part(content, cache_key))
-        elif msg.type == "function_call":
+        elif msg.type == "function_call" and not ignore_functions:
             parts.append(
                 types.Part(
                     function_call=types.FunctionCall(
@@ -68,7 +86,7 @@ def to_chat_ctx(
                     )
                 )
             )
-        elif msg.type == "function_call_output":
+        elif msg.type == "function_call_output" and not ignore_functions:
             parts.append(
                 types.Part(
                     function_response=types.FunctionResponse(

livekit/plugins/google/version.py

@@ -12,4 +12,4 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-__version__ = '1.0.0.rc9'
+__version__ = '1.0.2'

{livekit_plugins_google-1.0.0rc9.dist-info → livekit_plugins_google-1.0.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: livekit-plugins-google
-Version: 1.0.0rc9
+Version: 1.0.2
 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==1.5.0
-Requires-Dist: livekit-agents>=1.0.0.rc9
+Requires-Dist: livekit-agents>=1.0.2
 Description-Content-Type: text/markdown
 
 # LiveKit Plugins Google

livekit_plugins_google-1.0.2.dist-info/RECORD

@@ -0,0 +1,16 @@
+livekit/plugins/google/__init__.py,sha256=e_kSlFNmKhyyeliz7f4WOKc_Y0-y39QjO5nCWuguhss,1171
+livekit/plugins/google/llm.py,sha256=81LCCJPmpMOkApX0S0a-zu5xIvcm2Pk8lTTz-PoK5m0,14740
+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/utils.py,sha256=pbLSOAdQxInWhgI2Yhsrr9KvgvpFXYDdU2yx2p03pFg,9437
+livekit/plugins/google/version.py,sha256=VAosEGj0ByVVgOD0nuevp_anp63XZCFxkxz7t-41kg8,600
+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=ubF2Ha9zCD28gQrrjTcX3MWgMBs7bC3rI0DUdaHAa_Q,22021
+livekit_plugins_google-1.0.2.dist-info/METADATA,sha256=0sqwsTwIAhKGSWqP4QXvP4GX5LGcGxGIkN0AP6BDoNE,3489
+livekit_plugins_google-1.0.2.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+livekit_plugins_google-1.0.2.dist-info/RECORD,,

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

@@ -1,10 +0,0 @@
-def _build_tools():
-    pass
-
-
-def _create_ai_function_info():
-    pass
-
-
-def _build_gemini_ctx():
-    pass

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

@@ -1,254 +0,0 @@
-from __future__ import annotations
-
-import asyncio
-import re
-from dataclasses import dataclass
-from typing import Literal
-
-import websockets
-
-from google import genai
-from google.genai import types
-from google.genai.errors import APIError, ClientError, ServerError
-from livekit import rtc
-from livekit.agents import APIConnectionError, APIStatusError, utils
-
-from ...log import logger
-from .api_proto import ClientEvents, LiveAPIModels
-
-EventTypes = Literal["input_speech_started", "input_speech_done"]
-
-DEFAULT_LANGUAGE = "English"
-
-SYSTEM_INSTRUCTIONS = f"""
-You are an **Audio Transcriber**. Your task is to convert audio content into accurate and precise text.
-- Transcribe verbatim; exclude non-speech sounds.
-- Provide only transcription; no extra text or explanations.
-- If audio is unclear, respond with: `...`
-- Ensure error-free transcription, preserving meaning and context.
-- Use proper punctuation and formatting.
-- Do not add explanations, comments, or extra information.
-- Do not include timestamps, speaker labels, or annotations unless specified.
-- Audio Language: {DEFAULT_LANGUAGE}
-"""  # noqa: E501
-
-
-@dataclass
-class TranscriptionContent:
-    response_id: str
-    text: str
-
-
-class TranscriberSession(utils.EventEmitter[EventTypes]):
-    """
-    Handles live audio transcription using the realtime API.
-    """
-
-    def __init__(self, *, client: genai.Client, model: LiveAPIModels | str):
-        super().__init__()
-        self._client = client
-        self._model = model
-        self._needed_sr = 16000
-        self._closed = False
-
-        system_instructions = types.Content(parts=[types.Part(text=SYSTEM_INSTRUCTIONS)])
-        self._config = types.LiveConnectConfig(
-            response_modalities=[types.Modality.TEXT],
-            system_instruction=system_instructions,
-            generation_config=types.GenerationConfig(temperature=0.0),
-        )
-        self._main_atask = asyncio.create_task(
-            self._main_task(), name="gemini-realtime-transcriber"
-        )
-        self._send_ch = utils.aio.Chan[ClientEvents]()
-        self._resampler: rtc.AudioResampler | None = None
-        self._active_response_id = None
-
-    def _push_audio(self, frame: rtc.AudioFrame) -> None:
-        if self._closed:
-            return
-        if frame.sample_rate != self._needed_sr:
-            if not self._resampler:
-                self._resampler = rtc.AudioResampler(
-                    frame.sample_rate,
-                    self._needed_sr,
-                    quality=rtc.AudioResamplerQuality.HIGH,
-                )
-
-        if self._resampler:
-            for f in self._resampler.push(frame):
-                self._queue_msg(
-                    types.LiveClientRealtimeInput(
-                        media_chunks=[types.Blob(data=f.data.tobytes(), mime_type="audio/pcm")]
-                    )
-                )
-        else:
-            self._queue_msg(
-                types.LiveClientRealtimeInput(
-                    media_chunks=[types.Blob(data=frame.data.tobytes(), mime_type="audio/pcm")]
-                )
-            )
-
-    def _queue_msg(self, msg: ClientEvents) -> None:
-        if not self._closed:
-            self._send_ch.send_nowait(msg)
-
-    async def aclose(self) -> None:
-        if self._send_ch.closed:
-            return
-        self._closed = True
-        self._send_ch.close()
-        await self._main_atask
-
-    @utils.log_exceptions(logger=logger)
-    async def _main_task(self):
-        @utils.log_exceptions(logger=logger)
-        async def _send_task():
-            try:
-                async for msg in self._send_ch:
-                    if self._closed:
-                        break
-                    await self._session.send(input=msg)
-            except websockets.exceptions.ConnectionClosedError as e:
-                logger.exception(f"Transcriber session closed in _send_task: {e}")
-                self._closed = True
-            except Exception as e:
-                logger.exception(f"Uncaught error in transcriber _send_task: {e}")
-                self._closed = True
-
-        @utils.log_exceptions(logger=logger)
-        async def _recv_task():
-            try:
-                while not self._closed:
-                    async for response in self._session.receive():
-                        if self._closed:
-                            break
-                        if self._active_response_id is None:
-                            self._active_response_id = utils.shortuuid()
-                            content = TranscriptionContent(
-                                response_id=self._active_response_id,
-                                text="",
-                            )
-                            self.emit("input_speech_started", content)
-
-                        server_content = response.server_content
-                        if server_content:
-                            model_turn = server_content.model_turn
-                            if model_turn:
-                                for part in model_turn.parts:
-                                    if part.text:
-                                        content.text += part.text
-
-                            if server_content.turn_complete:
-                                content.text = clean_transcription(content.text)
-                                self.emit("input_speech_done", content)
-                                self._active_response_id = None
-
-            except websockets.exceptions.ConnectionClosedError as e:
-                logger.exception(f"Transcriber session closed in _recv_task: {e}")
-                self._closed = True
-            except Exception as e:
-                logger.exception(f"Uncaught error in transcriber _recv_task: {e}")
-                self._closed = True
-
-        async with self._client.aio.live.connect(model=self._model, config=self._config) as session:
-            self._session = session
-            tasks = [
-                asyncio.create_task(_send_task(), name="gemini-realtime-transcriber-send"),
-                asyncio.create_task(_recv_task(), name="gemini-realtime-transcriber-recv"),
-            ]
-
-            try:
-                await asyncio.gather(*tasks)
-            finally:
-                await utils.aio.gracefully_cancel(*tasks)
-                await self._session.close()
-
-
-class ModelTranscriber(utils.EventEmitter[EventTypes]):
-    """
-    Transcribes agent audio using model generation.
-    """
-
-    def __init__(self, *, client: genai.Client, model: LiveAPIModels | str):
-        super().__init__()
-        self._client = client
-        self._model = model
-        self._needed_sr = 16000
-        self._system_instructions = types.Content(parts=[types.Part(text=SYSTEM_INSTRUCTIONS)])
-        self._config = types.GenerateContentConfig(
-            temperature=0.0,
-            system_instruction=self._system_instructions,
-            # TODO: add response_schem
-        )
-        self._resampler: rtc.AudioResampler | None = None
-        self._buffer: rtc.AudioFrame | None = None
-        self._audio_ch = utils.aio.Chan[rtc.AudioFrame]()
-        self._main_atask = asyncio.create_task(self._main_task(), name="gemini-model-transcriber")
-
-    async def aclose(self) -> None:
-        if self._audio_ch.closed:
-            return
-        self._audio_ch.close()
-        await self._main_atask
-
-    def _push_audio(self, frames: list[rtc.AudioFrame]) -> None:
-        if not frames:
-            return
-
-        buffer = utils.merge_frames(frames)
-
-        if buffer.sample_rate != self._needed_sr:
-            if self._resampler is None:
-                self._resampler = rtc.AudioResampler(
-                    input_rate=buffer.sample_rate,
-                    output_rate=self._needed_sr,
-                    quality=rtc.AudioResamplerQuality.HIGH,
-                )
-
-            buffer = utils.merge_frames(self._resampler.push(buffer))
-
-        self._audio_ch.send_nowait(buffer)
-
-    @utils.log_exceptions(logger=logger)
-    async def _main_task(self):
-        request_id = utils.shortuuid()
-        try:
-            async for buffer in self._audio_ch:
-                # TODO: stream content for better latency
-                response = await self._client.aio.models.generate_content(
-                    model=self._model,
-                    contents=[
-                        types.Content(
-                            parts=[
-                                types.Part(text=SYSTEM_INSTRUCTIONS),
-                                types.Part.from_bytes(
-                                    data=buffer.to_wav_bytes(),
-                                    mime_type="audio/wav",
-                                ),
-                            ],
-                            role="user",
-                        )
-                    ],
-                    config=self._config,
-                )
-                content = TranscriptionContent(
-                    response_id=request_id, text=clean_transcription(response.text)
-                )
-                self.emit("input_speech_done", content)
-
-        except (ClientError, ServerError, APIError) as e:
-            raise APIStatusError(
-                f"model transcriber error: {e}",
-                status_code=e.code,
-                body=e.message,
-                request_id=request_id,
-            ) from e
-        except Exception as e:
-            raise APIConnectionError("Error generating transcription") from e
-
-
-def clean_transcription(text: str) -> str:
-    text = text.replace("\n", " ")
-    text = re.sub(r"\s+", " ", text)
-    return text.strip()

livekit_plugins_google-1.0.0rc9.dist-info/RECORD

@@ -1,18 +0,0 @@
-livekit/plugins/google/__init__.py,sha256=e_kSlFNmKhyyeliz7f4WOKc_Y0-y39QjO5nCWuguhss,1171
-livekit/plugins/google/llm.py,sha256=81LCCJPmpMOkApX0S0a-zu5xIvcm2Pk8lTTz-PoK5m0,14740
-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=fT5JtDM8ck2iMAzRvFKVeyT1oPt_R-bDkqiLa-ysikc,22539
-livekit/plugins/google/tts.py,sha256=aA3VuNaMcE6I1M43Sm-2mmvNyA9D2EyqfpyAporMUSg,8042
-livekit/plugins/google/utils.py,sha256=dLkq-8lbWDC7AQ7nULd9unWwu_Wv9czdlxchyiJQ2KQ,8740
-livekit/plugins/google/version.py,sha256=5bk2f3atP67YoCoyxLdm3aJrB_QkLXroUSkhYmjhT1o,604
-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=VO6QqOGOrxzsaOLBqnwNd8c-BId0PjwKicdrPTJisy0,688
-livekit/plugins/google/beta/realtime/realtime_api.py,sha256=ERM6WvcTtrfIyKpukzoSYrkhd3eYxIY-I09mKWp8vLk,22576
-livekit/plugins/google/beta/realtime/temp.py,sha256=an_YueuS_tUw3_QC6xWkkcw5JrJOBQFv2pJh6atpNcc,108
-livekit/plugins/google/beta/realtime/transcriber.py,sha256=DD7q894xc25GeeuKDar6-GwH-MxStEwhwBiX-KZ-Jo4,9559
-livekit_plugins_google-1.0.0rc9.dist-info/METADATA,sha256=l6sYTKKhQp1_JY5NwEZI5aRe8WvXfhgPx0gDMyicgJg,3496
-livekit_plugins_google-1.0.0rc9.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-livekit_plugins_google-1.0.0rc9.dist-info/RECORD,,