livekit-plugins-aws 1.2.2__tar.gz → 1.2.4__tar.gz

{livekit_plugins_aws-1.2.2 → livekit_plugins_aws-1.2.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: livekit-plugins-aws
-Version: 1.2.2
+Version: 1.2.4
 Summary: LiveKit Agents Plugin for services from AWS
 Project-URL: Documentation, https://docs.livekit.io
 Project-URL: Website, https://livekit.io/
@@ -20,7 +20,7 @@ Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Requires-Python: >=3.9.0
 Requires-Dist: aioboto3>=14.1.0
 Requires-Dist: amazon-transcribe>=0.6.2
-Requires-Dist: livekit-agents>=1.2.2
+Requires-Dist: livekit-agents>=1.2.4
 Provides-Extra: realtime
 Requires-Dist: aws-sdk-bedrock-runtime==0.0.2; (python_version >= '3.12') and extra == 'realtime'
 Requires-Dist: boto3>1.35.10; extra == 'realtime'

{livekit_plugins_aws-1.2.2 → livekit_plugins_aws-1.2.4}/livekit/plugins/aws/experimental/realtime/events.py

@@ -10,7 +10,19 @@ from ...log import logger
 
 MEDIA_TYPE = Literal["text/plain", "audio/lpcm", "application/json"]
 TYPE = Literal["TEXT", "AUDIO", "TOOL"]
-VOICE_ID = Literal["matthew", "tiffany", "amy"]
+VOICE_ID = Literal[
+    "matthew",
+    "tiffany",
+    "amy",
+    "ambre",
+    "florian",
+    "beatrice",
+    "lorenzo",
+    "greta",
+    "lennart",
+    "lupe",
+    "carlos",
+]
 ROLE = Literal["USER", "ASSISTANT", "TOOL", "SYSTEM"]
 GENERATION_STAGE = Literal["SPECULATIVE", "FINAL"]
 STOP_REASON = Literal["PARTIAL_TURN", "END_TURN", "INTERRUPTED"]

{livekit_plugins_aws-1.2.2 → livekit_plugins_aws-1.2.4}/livekit/plugins/aws/experimental/realtime/realtime_model.py

@@ -592,11 +592,16 @@ class RealtimeSession(  # noqa: F811
                 text_ch=utils.aio.Chan(),
                 audio_ch=utils.aio.Chan(),
             )
+            msg_modalities = asyncio.Future[list[Literal["text", "audio"]]]()
+            msg_modalities.set_result(
+                ["audio", "text"] if self._realtime_model.capabilities.audio_output else ["text"]
+            )
             self._current_generation.message_ch.send_nowait(
                 llm.MessageGeneration(
                     message_id=msg_gen.message_id,
                     text_stream=msg_gen.text_ch,
                     audio_stream=msg_gen.audio_ch,
+                    modalities=msg_modalities,
                 )
             )
             self._current_generation.messages[self._current_generation.response_id] = msg_gen
@@ -761,11 +766,16 @@ class RealtimeSession(  # noqa: F811
                 audio_ch=utils.aio.Chan(),
             )
             self._current_generation.messages[self._current_generation.response_id] = msg_gen
+            msg_modalities = asyncio.Future[list[Literal["text", "audio"]]]()
+            msg_modalities.set_result(
+                ["audio", "text"] if self._realtime_model.capabilities.audio_output else ["text"]
+            )
             self._current_generation.message_ch.send_nowait(
                 llm.MessageGeneration(
                     message_id=msg_gen.message_id,
                     text_stream=msg_gen.text_ch,
                     audio_stream=msg_gen.audio_ch,
+                    modalities=msg_modalities,
                 )
             )
             self.emit_generation_event()
@@ -1235,7 +1245,12 @@ class RealtimeSession(  # noqa: F811
         logger.warning("interrupt is not supported by Nova Sonic's Realtime API")
 
     def truncate(
-        self, *, message_id: str, audio_end_ms: int, audio_transcript: NotGivenOr[str] = NOT_GIVEN
+        self,
+        *,
+        message_id: str,
+        modalities: list[Literal["text", "audio"]],
+        audio_end_ms: int,
+        audio_transcript: NotGivenOr[str] = NOT_GIVEN,
     ) -> None:
         logger.warning("truncate is not supported by Nova Sonic's Realtime API")
 

{livekit_plugins_aws-1.2.2 → livekit_plugins_aws-1.2.4}/livekit/plugins/aws/stt.py

@@ -50,6 +50,7 @@ class STTOptions:
     enable_partial_results_stabilization: NotGivenOr[bool]
     partial_results_stability: NotGivenOr[str]
     language_model_name: NotGivenOr[str]
+    region: str
 
 
 class STT(stt.STT):
@@ -75,11 +76,6 @@ class STT(stt.STT):
 
         if not is_given(region):
             region = os.getenv("AWS_REGION") or DEFAULT_REGION
-        self._region = region
-        self._client = TranscribeStreamingClient(
-            region=self._region,
-            credential_resolver=AwsCrtCredentialResolver(None),  # type: ignore
-        )
 
         self._config = STTOptions(
             language=language,
@@ -95,6 +91,7 @@ class STT(stt.STT):
             enable_partial_results_stabilization=enable_partial_results_stabilization,
             partial_results_stability=partial_results_stability,
             language_model_name=language_model_name,
+            region=region,
         )
 
     async def aclose(self) -> None:
@@ -115,12 +112,7 @@ class STT(stt.STT):
         language: NotGivenOr[str] = NOT_GIVEN,
         conn_options: APIConnectOptions = DEFAULT_API_CONNECT_OPTIONS,
     ) -> SpeechStream:
-        return SpeechStream(
-            stt=self,
-            client=self._client,
-            conn_options=conn_options,
-            opts=self._config,
-        )
+        return SpeechStream(stt=self, conn_options=conn_options, opts=self._config)
 
 
 class SpeechStream(stt.SpeechStream):
@@ -128,15 +120,18 @@ class SpeechStream(stt.SpeechStream):
         self,
         stt: STT,
         opts: STTOptions,
-        client: TranscribeStreamingClient,
         conn_options: APIConnectOptions = DEFAULT_API_CONNECT_OPTIONS,
     ) -> None:
         super().__init__(stt=stt, conn_options=conn_options, sample_rate=opts.sample_rate)
         self._opts = opts
-        self._client = client
 
     async def _run(self) -> None:
         while True:
+            client = TranscribeStreamingClient(
+                region=self._opts.region,
+                credential_resolver=AwsCrtCredentialResolver(None),  # type: ignore
+            )
+
             live_config = {
                 "language_code": self._opts.language,
                 "media_sample_rate_hz": self._opts.sample_rate,
@@ -153,7 +148,7 @@ class SpeechStream(stt.SpeechStream):
                 "language_model_name": self._opts.language_model_name,
             }
             filtered_config = {k: v for k, v in live_config.items() if v and is_given(v)}
-            stream = await self._client.start_stream_transcription(**filtered_config)  # type: ignore
+            stream = await client.start_stream_transcription(**filtered_config)  # type: ignore
 
             async def input_generator(stream: StartStreamTranscriptionEventStream) -> None:
                 async for frame in self._input_ch:

{livekit_plugins_aws-1.2.2 → livekit_plugins_aws-1.2.4}/livekit/plugins/aws/version.py

@@ -12,4 +12,4 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-__version__ = "1.2.2"
+__version__ = "1.2.4"

{livekit_plugins_aws-1.2.2 → livekit_plugins_aws-1.2.4}/pyproject.toml

@@ -23,7 +23,7 @@ classifiers = [
     "Programming Language :: Python :: 3 :: Only",
 ]
 dependencies = [
-    "livekit-agents>=1.2.2",
+    "livekit-agents>=1.2.4",
     "aioboto3>=14.1.0",
     "amazon-transcribe>=0.6.2",
 ]