livekit-plugins-google 1.3.8__py3-none-any.whl → 1.3.11__py3-none-any.whl

livekit/plugins/google/tools.py

@@ -1,11 +1,71 @@
-from typing import Union
+from abc import ABC, abstractmethod
+from dataclasses import dataclass
+from typing import Optional
 
-from google.genai.types import (
-    GoogleMaps,
-    GoogleSearch,
-    GoogleSearchRetrieval,
-    ToolCodeExecution,
-    UrlContext,
-)
+from google.genai import types
+from livekit.agents import llm
 
-_LLMTool = Union[GoogleSearchRetrieval, ToolCodeExecution, GoogleSearch, UrlContext, GoogleMaps]
+
+class GeminiTool(llm.ProviderTool, ABC):
+    @abstractmethod
+    def to_tool_config(self) -> types.Tool: ...
+
+
+@dataclass
+class GoogleSearch(GeminiTool):
+    exclude_domains: Optional[list[str]] = None
+    blocking_confidence: Optional[types.PhishBlockThreshold] = None
+    time_range_filter: Optional[types.Interval] = None
+
+    def to_tool_config(self) -> types.Tool:
+        return types.Tool(
+            google_search=types.GoogleSearch(
+                exclude_domains=self.exclude_domains,
+                blocking_confidence=self.blocking_confidence,
+                time_range_filter=self.time_range_filter,
+            )
+        )
+
+
+@dataclass
+class GoogleMaps(GeminiTool):
+    auth_config: Optional[types.AuthConfig] = None
+    enable_widget: Optional[bool] = None
+
+    def to_tool_config(self) -> types.Tool:
+        return types.Tool(
+            google_maps=types.GoogleMaps(
+                auth_config=self.auth_config,
+                enable_widget=self.enable_widget,
+            )
+        )
+
+
+class URLContext(GeminiTool):
+    def to_tool_config(self) -> types.Tool:
+        return types.Tool(
+            url_context=types.UrlContext(),
+        )
+
+
+@dataclass
+class FileSearch(GeminiTool):
+    file_search_store_names: list[str]
+    top_k: Optional[int] = None
+    metadata_filter: Optional[str] = None
+
+    def to_tool_config(self) -> types.Tool:
+        return types.Tool(
+            file_search=types.FileSearch(
+                file_search_store_names=self.file_search_store_names,
+                top_k=self.top_k,
+                metadata_filter=self.metadata_filter,
+            )
+        )
+
+
+class ToolCodeExecution(GeminiTool):
+    def to_tool_config(self) -> types.Tool:
+        return types.Tool(
+            code_execution=types.ToolCodeExecution(),
+        )

livekit/plugins/google/tts.py

@@ -32,10 +32,9 @@ from livekit.agents.types import DEFAULT_API_CONNECT_OPTIONS, NOT_GIVEN, NotGive
 from livekit.agents.utils import is_given
 
 from .log import logger
-from .models import Gender, SpeechLanguages
+from .models import GeminiTTSModels, Gender, SpeechLanguages
 
 NUM_CHANNELS = 1
-DEFAULT_VOICE_NAME = "en-US-Chirp3-HD-Charon"
 DEFAULT_LANGUAGE = "en-US"
 DEFAULT_GENDER = "neutral"
 
@@ -65,7 +64,7 @@ class TTS(tts.TTS):
         gender: NotGivenOr[Gender | str] = NOT_GIVEN,
         voice_name: NotGivenOr[str] = NOT_GIVEN,
         voice_cloning_key: NotGivenOr[str] = NOT_GIVEN,
-        model_name: NotGivenOr[str] = NOT_GIVEN,
+        model_name: GeminiTTSModels | str = "gemini-2.5-flash-tts",
         prompt: NotGivenOr[str] = NOT_GIVEN,
         sample_rate: int = 24000,
         pitch: int = 0,
@@ -73,7 +72,7 @@ class TTS(tts.TTS):
         speaking_rate: float = 1.0,
         volume_gain_db: float = 0.0,
         location: str = "global",
-        audio_encoding: texttospeech.AudioEncoding = texttospeech.AudioEncoding.OGG_OPUS,  # type: ignore
+        audio_encoding: texttospeech.AudioEncoding = texttospeech.AudioEncoding.PCM,  # type: ignore
         credentials_info: NotGivenOr[dict] = NOT_GIVEN,
         credentials_file: NotGivenOr[str] = NOT_GIVEN,
         tokenizer: NotGivenOr[tokenize.SentenceTokenizer] = NOT_GIVEN,
@@ -92,9 +91,9 @@ class TTS(tts.TTS):
         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.
+            voice_name (str, optional): Specific voice name. Default is an empty string. See https://docs.cloud.google.com/text-to-speech/docs/gemini-tts#voice_options for supported voice in Gemini TTS models.
             voice_cloning_key (str, optional): Voice clone key. Created via https://cloud.google.com/text-to-speech/docs/chirp3-instant-custom-voice
-            model_name (str, optional): Model name for TTS (e.g., "gemini-2.5-flash-tts"). Enables Gemini TTS models with streaming support.
+            model_name (GeminiTTSModels | str, optional): Model name for TTS (e.g., "gemini-2.5-flash-tts", "chirp_3"). Default is "gemini-2.5-flash-tts".
             prompt (str, optional): Style prompt for Gemini TTS models. Controls tone, style, and speaking characteristics. Only applied to first input chunk in streaming mode.
             sample_rate (int, optional): Audio sample rate in Hz. Default is 24000.
             location (str, optional): Location for the TTS client. Default is "global".
@@ -134,14 +133,20 @@ class TTS(tts.TTS):
             language_code=lang,
             ssml_gender=ssml_gender,
         )
-        if is_given(model_name):
+        if model_name != "chirp_3":  #  voice_params.model_name must not be set for Chirp 3
             voice_params.model_name = model_name
+
         if is_given(voice_cloning_key):
             voice_params.voice_clone = texttospeech.VoiceCloneParams(
                 voice_cloning_key=voice_cloning_key,
             )
         else:
-            voice_params.name = voice_name if is_given(voice_name) else DEFAULT_VOICE_NAME
+            if is_given(voice_name):
+                voice_params.name = voice_name
+            elif model_name == "chirp_3":
+                voice_params.name = "en-US-Chirp3-HD-Charon"
+            else:
+                voice_params.name = "Charon"
 
         if not is_given(tokenizer):
             tokenizer = tokenize.blingfire.SentenceTokenizer()
@@ -160,7 +165,7 @@ class TTS(tts.TTS):
             custom_pronunciations=pronunciations,
             enable_ssml=enable_ssml,
             use_markup=use_markup,
-            model_name=model_name if is_given(model_name) else None,
+            model_name=model_name,
             prompt=prompt if is_given(prompt) else None,
         )
         self._streams = weakref.WeakSet[SynthesizeStream]()
@@ -285,6 +290,9 @@ class ChunkedStream(tts.ChunkedStream):
                     text=self._input_text, custom_pronunciations=self._opts.custom_pronunciations
                 )
 
+            if self._opts.prompt is not None:
+                tts_input.prompt = self._opts.prompt
+
             response: SynthesizeSpeechResponse = await self._tts._ensure_client().synthesize_speech(
                 input=tts_input,
                 voice=self._opts.voice,

livekit/plugins/google/utils.py

@@ -9,90 +9,40 @@ from pydantic import TypeAdapter
 from google.genai import types
 from livekit.agents import llm
 from livekit.agents.llm import utils as llm_utils
-from livekit.agents.llm.tool_context import (
-    FunctionTool,
-    RawFunctionTool,
-    get_raw_function_info,
-    is_function_tool,
-    is_raw_function_tool,
-)
 from livekit.agents.types import NOT_GIVEN, NotGivenOr
 from livekit.agents.utils import is_given
 
-from .log import logger
-from .tools import _LLMTool
+from .tools import GeminiTool
 
-__all__ = ["to_fnc_ctx"]
-
-
-def to_fnc_ctx(
-    fncs: list[FunctionTool | RawFunctionTool],
-    *,
-    use_parameters_json_schema: bool = True,
-    tool_behavior: NotGivenOr[types.Behavior] = NOT_GIVEN,
-) -> list[types.FunctionDeclaration]:
-    tools: list[types.FunctionDeclaration] = []
-    for fnc in fncs:
-        if is_raw_function_tool(fnc):
-            info = get_raw_function_info(fnc)
-            fnc_kwargs = {
-                "name": info.name,
-                "description": info.raw_schema.get("description", ""),
-            }
-            if use_parameters_json_schema:
-                fnc_kwargs["parameters_json_schema"] = info.raw_schema.get("parameters", {})
-            else:
-                # https://github.com/googleapis/python-genai/issues/1147
-                fnc_kwargs["parameters"] = types.Schema.from_json_schema(
-                    json_schema=types.JSONSchema.model_validate(
-                        info.raw_schema.get("parameters", {})
-                    )
-                )
-
-            if is_given(tool_behavior):
-                fnc_kwargs["behavior"] = tool_behavior
-            tools.append(types.FunctionDeclaration(**fnc_kwargs))
-
-        elif is_function_tool(fnc):
-            tools.append(_build_gemini_fnc(fnc, tool_behavior=tool_behavior))
-
-    return tools
+__all__ = ["create_tools_config"]
 
 
 def create_tools_config(
+    tool_ctx: llm.ToolContext,
     *,
-    function_tools: list[types.FunctionDeclaration] | None = None,
-    gemini_tools: list[_LLMTool] | None = None,
+    tool_behavior: NotGivenOr[types.Behavior] = NOT_GIVEN,
+    _only_single_type: bool = False,
 ) -> list[types.Tool]:
-    tools: list[types.Tool] = []
+    gemini_tools: list[types.Tool] = []
 
+    function_tools = [
+        types.FunctionDeclaration.model_validate(schema)
+        for schema in tool_ctx.parse_function_tools(
+            "google", tool_behavior=tool_behavior.value if tool_behavior else None
+        )
+    ]
     if function_tools:
-        tools.append(types.Tool(function_declarations=function_tools))
-
-    if gemini_tools:
-        for tool in gemini_tools:
-            if isinstance(tool, types.GoogleSearchRetrieval):
-                tools.append(types.Tool(google_search_retrieval=tool))
-            elif isinstance(tool, types.ToolCodeExecution):
-                tools.append(types.Tool(code_execution=tool))
-            elif isinstance(tool, types.GoogleSearch):
-                tools.append(types.Tool(google_search=tool))
-            elif isinstance(tool, types.UrlContext):
-                tools.append(types.Tool(url_context=tool))
-            elif isinstance(tool, types.GoogleMaps):
-                tools.append(types.Tool(google_maps=tool))
-            else:
-                logger.warning(f"Warning: Received unhandled tool type: {type(tool)}")
-                continue
+        gemini_tools.append(types.Tool(function_declarations=function_tools))
 
-    if len(tools) > 1:
-        # https://github.com/google/adk-python/issues/53#issuecomment-2799538041
-        logger.warning(
-            "Multiple kinds of tools are not supported in Gemini. Only the first tool will be used."
-        )
-        tools = tools[:1]
+    # Some Google LLMs do not support multiple tool types (either function tools or builtin tools).
+    if _only_single_type and gemini_tools:
+        return gemini_tools
+
+    for tool in tool_ctx.provider_tools:
+        if isinstance(tool, GeminiTool):
+            gemini_tools.append(tool.to_tool_config())
 
-    return tools
+    return gemini_tools
 
 
 def get_tool_results_for_realtime(
@@ -124,22 +74,6 @@ def get_tool_results_for_realtime(
     )
 
 
-def _build_gemini_fnc(
-    function_tool: FunctionTool, *, tool_behavior: NotGivenOr[types.Behavior] = NOT_GIVEN
-) -> types.FunctionDeclaration:
-    fnc = llm.utils.build_legacy_openai_schema(function_tool, internally_tagged=True)
-    json_schema = _GeminiJsonSchema(fnc["parameters"]).simplify()
-
-    kwargs = {
-        "name": fnc["name"],
-        "description": fnc["description"],
-        "parameters": types.Schema.model_validate(json_schema) if json_schema else None,
-    }
-    if is_given(tool_behavior):
-        kwargs["behavior"] = tool_behavior
-    return types.FunctionDeclaration(**kwargs)
-
-
 def to_response_format(response_format: type | dict) -> types.SchemaUnion:
     _, json_schema_type = llm_utils.to_response_format_param(response_format)
     if isinstance(json_schema_type, TypeAdapter):

livekit/plugins/google/version.py

@@ -12,4 +12,4 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-__version__ = "1.3.8"
+__version__ = "1.3.11"

{livekit_plugins_google-1.3.8.dist-info → livekit_plugins_google-1.3.11.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: livekit-plugins-google
-Version: 1.3.8
+Version: 1.3.11
 Summary: Agent Framework plugin for services from Google Cloud
 Project-URL: Documentation, https://docs.livekit.io
 Project-URL: Website, https://livekit.io/

livekit_plugins_google-1.3.11.dist-info/RECORD

@@ -0,0 +1,18 @@
+livekit/plugins/google/__init__.py,sha256=21ZYfsz4d4a5tP_po9WPtIv552gYh6thg7mcnkYA9vc,1445
+livekit/plugins/google/llm.py,sha256=eLLuXQZ0GMrl8blfb2staHr2ClAsYroJRydt0arT1Uk,24386
+livekit/plugins/google/log.py,sha256=GI3YWN5YzrafnUccljzPRS_ZALkMNk1i21IRnTl2vNA,69
+livekit/plugins/google/models.py,sha256=NOkEJVTmvZ7A6TSVCACoaST-qM84YSPwo_HkM5ct1mY,3171
+livekit/plugins/google/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+livekit/plugins/google/stt.py,sha256=Q3-gEAV5xykTLfmkj8bPq6mHdx4cn6W7_-fQrI7n85o,35246
+livekit/plugins/google/tools.py,sha256=cH8qXkQj6zQ_cWUV-4apjSeQaMhfK6907XKPAs9PVNo,1956
+livekit/plugins/google/tts.py,sha256=dVhGDN0Q-b_tGKdridkb5grl8YhBMcxTc_kIdcg4gQ8,19251
+livekit/plugins/google/utils.py,sha256=MSMsmPBsmxL-rSi5mc8d7ViI5bHgNEOO4Em0yHLhNyQ,8128
+livekit/plugins/google/version.py,sha256=Db2pVVYNC02fj0G6tMcFzTjBbAxJWVcZ5l1gTIq9VU4,601
+livekit/plugins/google/beta/__init__.py,sha256=4q5dx-Y6o9peCDziB03Skf5ngH4PTBsZC86ZawWrgnk,271
+livekit/plugins/google/beta/gemini_tts.py,sha256=SpKorOteQ7GYoGWsxV5YPuGeMexoosmtDXQVz_1ZeLA,8743
+livekit/plugins/google/realtime/__init__.py,sha256=_fW2NMN22F-hnQ4xAJ_g5lPbR7CvM_xXzSWlUQY-E-U,188
+livekit/plugins/google/realtime/api_proto.py,sha256=n6Rb-3qrZyByp8MSkBHA3TIW2E0IGKH1Xj7hKJM029M,2290
+livekit/plugins/google/realtime/realtime_api.py,sha256=_eCRUK2Bi4ypBEPcCzK9ym_isO8BAvmlXkgW1cO4P8w,54723
+livekit_plugins_google-1.3.11.dist-info/METADATA,sha256=sQqXHcC_xM0n6_MNZYlJJq61WRG4y422eYFDy0YeXD4,2467
+livekit_plugins_google-1.3.11.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+livekit_plugins_google-1.3.11.dist-info/RECORD,,

livekit_plugins_google-1.3.8.dist-info/RECORD

@@ -1,18 +0,0 @@
-livekit/plugins/google/__init__.py,sha256=bYHN04-Ttynj09POAnFP3mln-wrEc1vanUD_YpoWOE4,1434
-livekit/plugins/google/llm.py,sha256=OSl-3DzYRGPUvmL4go-iCc_iag9woyvn5ZTpElLAnd0,21396
-livekit/plugins/google/log.py,sha256=GI3YWN5YzrafnUccljzPRS_ZALkMNk1i21IRnTl2vNA,69
-livekit/plugins/google/models.py,sha256=jsXHLSCDw-T5dZXeDE2nMT2lr0GooCYO4y4aW7Htps4,2816
-livekit/plugins/google/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-livekit/plugins/google/stt.py,sha256=fsWoNnpjgLxqY43cx6GbRI-_QLvXmMvD4WczJFjnoOA,26846
-livekit/plugins/google/tools.py,sha256=tD5HVDHO5JfUF029Cx3axHMJec0Gxalkl7s1FDgxLzI,259
-livekit/plugins/google/tts.py,sha256=6qQDLRawqW5SSoN4JAqZYjyANxKT3165zJDfO6_7f08,18846
-livekit/plugins/google/utils.py,sha256=U_h7kV3528uH-Ef946yaMjS-Y9rH_yreu32fQ6qwVNo,10657
-livekit/plugins/google/version.py,sha256=QU6KNV-p6CdJv8bUwDgVMAemqOx2EwSsOFOuYh4-Re0,600
-livekit/plugins/google/beta/__init__.py,sha256=4q5dx-Y6o9peCDziB03Skf5ngH4PTBsZC86ZawWrgnk,271
-livekit/plugins/google/beta/gemini_tts.py,sha256=SpKorOteQ7GYoGWsxV5YPuGeMexoosmtDXQVz_1ZeLA,8743
-livekit/plugins/google/realtime/__init__.py,sha256=_fW2NMN22F-hnQ4xAJ_g5lPbR7CvM_xXzSWlUQY-E-U,188
-livekit/plugins/google/realtime/api_proto.py,sha256=6zZ5foiBkHfq6p9-fPf-fzBACAKFj5G7tJ_2PWPpoLg,1432
-livekit/plugins/google/realtime/realtime_api.py,sha256=SMEJ6ENBwVBt5DapcIaTiwzY9ANbeMDv_5tjifiQ4Ec,55231
-livekit_plugins_google-1.3.8.dist-info/METADATA,sha256=HMmh2kVB5WDIAbGostE7cUIWJ-6-xE9tY6ipJrJ2rHo,2466
-livekit_plugins_google-1.3.8.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
-livekit_plugins_google-1.3.8.dist-info/RECORD,,