Merge branch 'main' into feat/async-tts

src/pipecat/processors/aggregators/openai_llm_context.py

@@ -111,8 +111,6 @@ class OpenAILLMContext:
         context = OpenAILLMContext()
 
         for message in messages:
-            if "name" not in message:
-                message["name"] = message["role"]
             context.add_message(message)
         return context
 

src/pipecat/processors/transcript_processor.py

@@ -140,11 +140,13 @@ class AssistantTranscriptProcessor(BaseTranscriptProcessor):
                 Result: "Hello there how are you"
         """
         if self._current_text_parts and self._aggregation_start_time:
+            # Check specifically for space characters, previously isspace() was used
+            # but that includes all whitespace characters (e.g. \n), not just spaces.
             has_leading_spaces = any(
-                part and part[0].isspace() for part in self._current_text_parts[1:]
+                part and part[0] == " " for part in self._current_text_parts[1:]
             )
             has_trailing_spaces = any(
-                part and part[-1].isspace() for part in self._current_text_parts[:-1]
+                part and part[-1] == " " for part in self._current_text_parts[:-1]
             )
 
             # If there are embedded spaces in the fragments, use direct concatenation

src/pipecat/transports/services/daily.py

@@ -226,6 +226,8 @@ class DailyCallbacks(BaseModel):
         on_participant_left: Called when a participant leaves.
         on_participant_updated: Called when participant info is updated.
         on_transcription_message: Called when receiving transcription.
+        on_transcription_stopped: Called when transcription is stopped.
+        on_transcription_error: Called when transcription encounters an error.
         on_recording_started: Called when recording starts.
         on_recording_stopped: Called when recording stops.
         on_recording_error: Called when recording encounters an error.
@@ -253,6 +255,8 @@ class DailyCallbacks(BaseModel):
     on_participant_left: Callable[[Mapping[str, Any], str], Awaitable[None]]
     on_participant_updated: Callable[[Mapping[str, Any]], Awaitable[None]]
     on_transcription_message: Callable[[Mapping[str, Any]], Awaitable[None]]
+    on_transcription_stopped: Callable[[str, bool], Awaitable[None]]
+    on_transcription_error: Callable[[str], Awaitable[None]]
     on_recording_started: Callable[[Mapping[str, Any]], Awaitable[None]]
     on_recording_stopped: Callable[[str], Awaitable[None]]
     on_recording_error: Callable[[str, str], Awaitable[None]]
@@ -1233,6 +1237,9 @@ class DailyTransportClient(EventHandler):
             stopped_by_error: Whether stopped due to error.
         """
         logger.debug("Transcription stopped")
+        self._call_event_callback(
+            self._callbacks.on_transcription_stopped, stopped_by, stopped_by_error
+        )
 
     def on_transcription_error(self, message):
         """Handle transcription error events.
@@ -1241,6 +1248,7 @@ class DailyTransportClient(EventHandler):
             message: Error message.
         """
         logger.error(f"Transcription error: {message}")
+        self._call_event_callback(self._callbacks.on_transcription_error, message)
 
     def on_transcription_message(self, message):
         """Handle transcription message events.
@@ -1834,6 +1842,8 @@ class DailyTransport(BaseTransport):
             on_participant_left=self._on_participant_left,
             on_participant_updated=self._on_participant_updated,
             on_transcription_message=self._on_transcription_message,
+            on_transcription_stopped=self._on_transcription_stopped,
+            on_transcription_error=self._on_transcription_error,
             on_recording_started=self._on_recording_started,
             on_recording_stopped=self._on_recording_stopped,
             on_recording_error=self._on_recording_error,
@@ -2318,6 +2328,14 @@ class DailyTransport(BaseTransport):
         if self._input:
             await self._input.push_transcription_frame(frame)
 
+    async def _on_transcription_stopped(self, stopped_by, stopped_by_error):
+        """Handle transcription stopped events."""
+        await self._call_event_handler("on_transcription_stopped", stopped_by, stopped_by_error)
+
+    async def _on_transcription_error(self, message):
+        """Handle transcription error events."""
+        await self._call_event_handler("on_transcription_error", message)
+
     async def _on_recording_started(self, status):
         """Handle recording started events."""
         await self._call_event_handler("on_recording_started", status)

src/pipecat/transports/services/helpers/daily_rest.py

@@ -52,6 +52,25 @@ class RecordingsBucketConfig(BaseModel):
     allow_api_access: bool = False
 
 
+class TranscriptionBucketConfig(BaseModel):
+    """Configuration for storing Daily transcription in a custom S3 bucket.
+
+    Refer to the Daily API documentation for more information:
+    https://docs.daily.co/guides/products/live-streaming-recording/storing-recordings-in-a-custom-s3-bucket
+
+    Parameters:
+        bucket_name: Name of the S3 bucket for storing transcription.
+        bucket_region: AWS region where the S3 bucket is located.
+        assume_role_arn: ARN of the IAM role to assume for S3 access.
+        allow_api_access: Whether to allow API access to the transcription.
+    """
+
+    bucket_name: str
+    bucket_region: str
+    assume_role_arn: str
+    allow_api_access: bool = False
+
+
 class DailyRoomProperties(BaseModel, extra="allow"):
     """Properties for configuring a Daily room.
 
@@ -65,9 +84,11 @@ class DailyRoomProperties(BaseModel, extra="allow"):
         eject_at_room_exp: Whether to remove participants when room expires.
         enable_dialout: Whether SIP dial-out is enabled.
         enable_recording: Recording settings ('cloud', 'local', 'raw-tracks').
+        enable_transcription_storage: Whether transcription storage is enabled.
         geo: Geographic region for room.
         max_participants: Maximum number of participants allowed in the room.
         recordings_bucket: Configuration for custom S3 bucket recordings.
+        transcription_bucket: Configuration for custom S3 bucket transcription.
         sip: SIP configuration parameters.
         sip_uri: SIP URI information returned by Daily.
         start_video_off: Whether video is off by default.
@@ -80,9 +101,11 @@ class DailyRoomProperties(BaseModel, extra="allow"):
     eject_at_room_exp: bool = False
     enable_dialout: Optional[bool] = None
     enable_recording: Optional[Literal["cloud", "local", "raw-tracks"]] = None
+    enable_transcription_storage: Optional[bool] = None
     geo: Optional[str] = None
     max_participants: Optional[int] = None
     recordings_bucket: Optional[RecordingsBucketConfig] = None
+    transcription_bucket: Optional[TranscriptionBucketConfig] = None
     sip: Optional[DailyRoomSipParams] = None
     sip_uri: Optional[dict] = None
     start_video_off: bool = False

src/pipecat/utils/asyncio/watchdog_priority_queue.py

@@ -63,7 +63,7 @@ class WatchdogPriorityQueue(asyncio.PriorityQueue):
             get_result = await super().get()
 
         if isinstance(get_result, _WatchdogPriorityCancelSentinel):
-            logger.debug(
+            logger.trace(
                 "Received WatchdogPriorityCancelSentinel, throwing CancelledError to force cancelling"
             )
             raise asyncio.CancelledError("Cancelling watchdog queue get() call.")

src/pipecat/utils/asyncio/watchdog_queue.py

@@ -62,7 +62,7 @@ class WatchdogQueue(asyncio.Queue):
             get_result = await super().get()
 
         if isinstance(get_result, _WatchdogQueueCancelSentinel):
-            logger.debug(
+            logger.trace(
                 "Received WatchdogQueueCancelFrame, throwing CancelledError to force cancelling"
             )
             raise asyncio.CancelledError("Cancelling watchdog queue get() call.")