more demo cleanup, allow bundled frames in output_queue

src/dailyai/services/daily_transport_service.py

@@ -147,7 +147,9 @@ class DailyTransportService(EventHandler):
 
     def stop(self):
         self.stop_threads.set()
+        if self.camera_thread and self.camera_thread.is_alive():
             self.camera_thread.join()
+        if self.frame_consumer_thread and self.frame_consumer_thread.is_alive():
             self.output_queue.put(OutputQueueFrame(FrameType.END_STREAM, None))
             self.frame_consumer_thread.join()
         self.client.leave()
@@ -223,7 +225,11 @@ class DailyTransportService(EventHandler):
         all_audio_frames = bytearray()
         while True:
             try:
-                frame: OutputQueueFrame = self.output_queue.get()
+                frames: OutputQueueFrame | list[OutputQueueFrame] = self.output_queue.get()
+                if type(frames) != list:
+                    frames = [frames]
+
+                for frame in frames:
                     if frame.frame_type == FrameType.END_STREAM:
                         self.logger.info("Stopping frame consumer thread")
                         return

src/samples/05-sync-speech-and-text.py

@@ -26,32 +26,52 @@ async def main(room_url, token):
     tts = AzureTTSService()
     dalle = AzureImageGenServiceREST()
 
-    inference_text_process = llm.run_llm(
+    async def get_all_audio(text):
+        all_audio = bytearray()
+        async for audio in tts.run_tts(text):
+            all_audio.append(audio)
+
+        return all_audio
+
+    async def show_month(month):
+        print(f"Running llm for {month}")
+        inference_text = await llm.run_llm(
             [
                 {
                     "role": "system",
-                "content": f"Describe a nature photograph suitable for use in a calendar, for the month of January. Include only the image description with no preamble."
+                    "content": f"Describe a nature photograph suitable for use in a calendar, for the month of {month}. Include only the image description with no preamble."
                 }
             ]
         )
+        print(f"got llm for {month}")
+
+        (image, audio) = await asyncio.gather(
+            *[dalle.run_image_gen(inference_text, "1024x1024"), get_all_audio(inference_text)]
+        )
+        print(f"Got audio and video for {month}")
+        transport.output_queue.put(
+            [
+                OutputQueueFrame(FrameType.IMAGE_FRAME, image[1]),
+                OutputQueueFrame(FrameType.AUDIO_FRAME, audio),
+            ]
+        )
 
     try:
         transport.run()
-
+        months = [
-        inference_text = await inference_text_process
+            "January",
-
+            "February",
-        tts_iterator = tts.run_tts(inference_text)
+            "March",
-        (image, audio) = await asyncio.gather(
+            "April",
-            *[dalle.run_image_gen(inference_text, "1024x1024"), anext(tts_iterator)]
+            "May",
-        )
+            "June",
-        transport.output_queue.put(OutputQueueFrame(FrameType.IMAGE_FRAME, image[1]))
+            "July",
-        transport.output_queue.put(OutputQueueFrame(FrameType.AUDIO_FRAME, audio))
+            "August",
-        async for audio in tts_iterator:
+            "September",
-            transport.output_queue.put(
+            "October",
-                OutputQueueFrame(FrameType.AUDIO_FRAME, audio)
+            "November",
-            )
+        ]
-
+        await asyncio.gather(*[show_month(month) for month in months])
-        await asyncio.sleep(meeting_duration_minutes * 60)
     finally:
         transport.stop()
     print("Done")