first version push to talk

agents/my_basic_agent_1_2_9.py

@@ -31,6 +31,7 @@ from livekit.agents import (
     cli,
     get_job_context,
     metrics,
+    RoomIO
 )
 from livekit.agents.llm import ImageContent, ToolError, function_tool
 from typing import Any, List, Optional
@@ -953,6 +954,8 @@ async def entrypoint(ctx: JobContext, avatar_dispatcher_url: str = None, vision_
         # Increase the maximum number of function calls per turn to avoid hitting the limit
         max_tool_steps=15,
     )
+    room_io = RoomIO(session, room=ctx.room)
+    await room_io.start()
 
     # log metrics as they are emitted, and total usage after session is over
     usage_collector = metrics.UsageCollector()
@@ -1011,6 +1014,45 @@ async def entrypoint(ctx: JobContext, avatar_dispatcher_url: str = None, vision_
         room_output_options=RoomOutputOptions(transcription_enabled=True),
     )
 
+     # disable input audio at the start
+    session.input.set_audio_enabled(False)
+
+    @ctx.room.local_participant.register_rpc_method("start_turn")
+    async def start_turn(data: rtc.RpcInvocationData):
+        try:
+            session.interrupt()
+        except RuntimeError as e:
+            logger.error(f"Failed to interrupt session: {e}")
+            # Raise RPC error so client can detect interrupt failure
+            # Use ERROR_INTERNAL (code 13) to indicate application error
+            raise rtc.RpcError(
+                code=13,  # ERROR_INTERNAL
+                message="Application error in method handler"
+            )
+
+        session.clear_user_turn()
+
+        # listen to the caller if multi-user
+        room_io.set_participant(data.caller_identity)
+        session.input.set_audio_enabled(True)
+
+    @ctx.room.local_participant.register_rpc_method("end_turn")
+    async def end_turn(data: rtc.RpcInvocationData):
+        session.input.set_audio_enabled(False)
+        session.commit_user_turn(
+            # the timeout for the final transcript to be received after committing the user turn
+            # increase this value if the STT is slow to respond
+            transcript_timeout=10.0,
+            # the duration of the silence to be appended to the STT to make it generate the final transcript
+            stt_flush_duration=2.0,
+        )
+
+    @ctx.room.local_participant.register_rpc_method("cancel_turn")
+    async def cancel_turn(data: rtc.RpcInvocationData):
+        session.input.set_audio_enabled(False)
+        session.clear_user_turn()
+        logger.info("cancel turn")
+
 if __name__ == "__main__":
     parser = argparse.ArgumentParser()
     parser.add_argument("--avatar-url", type=str, default=None, help="Avatar dispatcher URL")