livekit-plugins-volcengine 1.2.2__tar.gz → 1.2.3.post0__tar.gz

{livekit_plugins_volcengine-1.2.2 → livekit_plugins_volcengine-1.2.3.post0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: livekit-plugins-volcengine
-Version: 1.2.2
+Version: 1.2.3.post0
 Summary: LiveKit Agent Plugins for Volcengine
 Author-email: wangmengdi <790990241@qq.com>
 Keywords: audio,livekit,realtime,video,webrtc
@@ -14,7 +14,7 @@ Classifier: Topic :: Multimedia :: Sound/Audio
 Classifier: Topic :: Multimedia :: Video
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Requires-Python: >=3.9
-Requires-Dist: livekit-agents>=1.2.2
+Requires-Dist: livekit-agents==1.2.2
 Requires-Dist: numpy
 Requires-Dist: openai>=1.75.0
 Requires-Dist: osc-data>=0.2.2

{livekit_plugins_volcengine-1.2.2 → livekit_plugins_volcengine-1.2.3.post0}/livekit/plugins/volcengine/realtime.py

@@ -171,6 +171,7 @@ class _RealtimeOptions:
     conn_options: APIConnectOptions
     opening: str = "你好啊，今天过得怎么样？"
     speaking_style: str = "你的说话风格简洁明了，语速适中，语调自然。"
+    speaker: str = "zh_female_vv_jupiter_bigtts"
     sample_rate: int = 24000
     num_channels: int = 1
     format: str = "pcm"
@@ -197,6 +198,7 @@ class _RealtimeOptions:
                     "format": self.format,
                     "sample_rate": self.sample_rate,
                 },
+                "speaker": self.speaker,
             },
             "dialog": {
                 "bot_name": self.bot_name,
@@ -214,6 +216,7 @@ class _MessageGeneration:
     message_id: str
     text_ch: utils.aio.Chan[str]
     audio_ch: utils.aio.Chan[rtc.AudioFrame]
+    modalities: asyncio.Future[list[Literal["text", "audio"]]]
     audio_transcript: str = ""
 
 
@@ -236,6 +239,7 @@ class RealtimeModel(llm.RealtimeModel):
         self,
         bot_name: str = "豆包",
         speaking_style: str = "你的说话风格简洁明了，语速适中，语调自然。",
+        speaker: str = "zh_female_vv_jupiter_bigtts",
         opening: str | None = None,
         app_id: str | None = None,
         access_token: str | None = None,
@@ -270,11 +274,13 @@ class RealtimeModel(llm.RealtimeModel):
             conn_options=conn_options,
             bot_name=bot_name,
             system_role=system_role,
+            speaker=speaker,
             opening=opening,
             speaking_style=speaking_style,
         )
         self._http_session = http_session
         self._sessions = weakref.WeakSet[RealtimeSession]()
+        print(self.capabilities.audio_output)
 
     def update_options(
         self,
@@ -436,13 +442,18 @@ class RealtimeSession(
                 message_id=item_id,
                 text_ch=utils.aio.Chan(),
                 audio_ch=utils.aio.Chan(),
+                modalities=asyncio.Future(),
             )
+            if not self._realtime_model.capabilities.audio_output:
+                self._current_item.audio_ch.close()
+                self._current_item.modalities.set_result(["text"])
 
             self._current_generation.message_ch.send_nowait(
                 llm.MessageGeneration(
                     message_id=item_id,
                     text_stream=self._current_item.text_ch,
                     audio_stream=self._current_item.audio_ch,
+                    modalities=self._current_item.modalities,
                 )
             )
 
@@ -494,6 +505,7 @@ class RealtimeSession(
                         is_final = not response["results"][0]["is_interim"]
                         if is_final:
                             item_id = utils.shortuuid()
+                            logger.info("transcription completed")
                             self.emit(
                                 "input_audio_transcription_completed",
                                 llm.InputTranscriptionCompleted(
@@ -523,12 +535,17 @@ class RealtimeSession(
                                     message_id=item_id,
                                     text_ch=utils.aio.Chan(),
                                     audio_ch=utils.aio.Chan(),
+                                    modalities=asyncio.Future(),
                                 )
+                                if not self._realtime_model.capabilities.audio_output:
+                                    self._current_item.audio_ch.close()
+                                    self._current_item.modalities.set_result(["text"])
                                 self._current_generation.message_ch.send_nowait(
                                     llm.MessageGeneration(
                                         message_id=item_id,
                                         text_stream=self._current_item.text_ch,
                                         audio_stream=self._current_item.audio_ch,
+                                        modalities=self._current_item.modalities,
                                     )
                                 )
 
@@ -544,6 +561,7 @@ class RealtimeSession(
                         logger.info("tts start")
 
                     elif event == 352:  # TTSResponse
+                        logger.info("tts response")
                         if self._first_tts_response:
                             logger.info("llm first sentence")
                             logger.info("tts first response")

livekit_plugins_volcengine-1.2.3.post0/livekit/plugins/volcengine/version.py

@@ -0,0 +1 @@
+__version__ = "1.2.3.post0"

{livekit_plugins_volcengine-1.2.2 → livekit_plugins_volcengine-1.2.3.post0}/pyproject.toml

@@ -9,11 +9,11 @@ authors = [
 keywords = ["webrtc", "realtime", "audio", "video", "livekit"]
 requires-python = ">=3.9"
 dependencies = [
-    "livekit-agents>=1.2.2",
     "openai>=1.75.0",
     "osc-data>=0.2.2",
     "pydantic",
     "numpy",
+    "livekit-agents==1.2.2",
 ]
 
 classifiers = [

livekit_plugins_volcengine-1.2.2/livekit/plugins/volcengine/version.py

@@ -1 +0,0 @@
-__version__ = "1.2.2"