little more cleanup

src/dailyai/conversation_wrappers.py

@@ -2,7 +2,7 @@ import asyncio
 import copy
 import functools
 from typing import AsyncGenerator, Awaitable, Callable
-from dailyai.queue_aggregators import LLMContextAggregator
+from dailyai.queue_aggregators import LLMAssistantContextAggregator, LLMContextAggregator, LLMUserContextAggregator
 from dailyai.queue_frame import EndStreamQueueFrame, QueueFrame, TranscriptionQueueFrame
 
 
@@ -17,8 +17,8 @@ class InterruptibleConversationWrapper:
         interrupt: Callable[[], None],
         my_participant_id: str | None,
         llm_messages: list[dict[str, str]],
-        llm_context_aggregator_in=LLMContextAggregator,
-        llm_context_aggregator_out=LLMContextAggregator,
+        llm_context_aggregator_in=LLMUserContextAggregator,
+        llm_context_aggregator_out=LLMAssistantContextAggregator,
         delay_before_speech_seconds: float = 1.0,
     ):
         self._frame_generator: Callable[[], AsyncGenerator[QueueFrame, None]] = frame_generator

src/dailyai/queue_aggregators.py

@@ -33,7 +33,7 @@ class LLMContextAggregator(AIService):
             role: str,
             bot_participant_id=None,
             complete_sentences=True,
-            pass_through=False):
+            pass_through=True):
         self.messages = messages
         self.bot_participant_id = bot_participant_id
         self.role = role
@@ -42,28 +42,32 @@ class LLMContextAggregator(AIService):
         self.pass_through = pass_through
 
     async def process_frame(self, frame: QueueFrame) -> AsyncGenerator[QueueFrame, None]:
-        # TODO: split up transcription by participant
-        if isinstance(frame, TextQueueFrame):
-
-            # Ignore transcription frames from the bot
-            if isinstance(frame, TranscriptionQueueFrame):
-                if frame.participantId == self.bot_participant_id:
-                    return
-
-            if self.complete_sentences:
-                self.sentence += frame.text
-                if self.sentence.endswith((".", "?", "!")):
-                    self.messages.append({"role": self.role, "content": self.sentence})
-                    self.sentence = ""
-                    yield LLMMessagesQueueFrame(self.messages)
-            else:
-                self.messages.append({"role": self.role, "content": frame.text})
-                yield LLMMessagesQueueFrame(self.messages)
-
-            if self.pass_through:
-                yield frame
-        else:
+        # We don't do anything with non-text frames, pass it along to next in the pipeline.
+        if not isinstance(frame, TextQueueFrame):
             yield frame
+            return
+
+        # The common case for "pass through" is receiving frames from the LLM that we'll
+        # use to update the "assistant" LLM messages, but also passing the text frames
+        # along to a TTS service to be spoken to the user.
+        if self.pass_through:
+            yield frame
+
+        # Ignore transcription frames from the bot
+        if isinstance(frame, TranscriptionQueueFrame):
+            if frame.participantId == self.bot_participant_id:
+                return
+
+        # TODO: split up transcription by participant
+        if self.complete_sentences:
+            self.sentence += frame.text # type: ignore -- the linter thinks this isn't a TextQueueFrame, even though we check it above
+            if self.sentence.endswith((".", "?", "!")):
+                self.messages.append({"role": self.role, "content": self.sentence})
+                self.sentence = ""
+                yield LLMMessagesQueueFrame(self.messages)
+        else:
+            self.messages.append({"role": self.role, "content": frame.text})  # type: ignore -- the linter thinks this isn't a TextQueueFrame, even though we check it above
+            yield LLMMessagesQueueFrame(self.messages)
 
 class LLMUserContextAggregator(LLMContextAggregator):
     def __init__(self,

src/samples/foundational/07-interruptible.py

@@ -25,6 +25,7 @@ async def main(room_url: str, token):
     transport.mic_enabled = True
     transport.mic_sample_rate = 16000
     transport.camera_enabled = False
+    transport.start_transcription = True
 
     llm = AzureLLMService()
     tts = ElevenLabsTTSService(voice_id="ErXwobaYiN019PkySvjV")