wx44wx/pipecat
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

refactor party tonight

Browse Source
This commit is contained in:
Moishe Lettvin
2024-01-17 18:42:08 -05:00
parent a3ac0d84e8
commit 13f2f792af
10 changed files with 187 additions and 118 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
src/samples/theoretical-to-real/01-say-one-thing.py
View File

@@ -27,21 +27,16 @@ async def main(room_url):
# similarly, create a tts service
tts = AzureTTSService()
# Get the generator for the audio. This will start running in the background,
# and when we ask the generator for its items, we'll get what it's generated.
audio_generator: AsyncGenerator[bytes, None] = tts.run_tts("hello world")
# Register an event handler so we can play the audio when the participant joins.
@transport.event_handler("on_participant_joined")
async def on_participant_joined(transport, participant):
if participant["info"]["isLocal"]:
return
async for audio in audio_generator:
transport.output_queue.put(QueueFrame(FrameType.AUDIO, audio))
await tts.say("Hello there, " + participant["info"]["userName"] + "!", transport.send_queue)
# wait for the output queue to be empty, then leave the meeting
transport.output_queue.join()
transport.wait_for_send_queue_to_empty()
transport.stop()
await transport.run()

27
src/samples/theoretical-to-real/02-llm-say-one-thing.py
View File

@@ -4,6 +4,7 @@ from typing import AsyncGenerator
from dailyai.queue_frame import QueueFrame, FrameType
from dailyai.services.daily_transport_service import DailyTransportService
from dailyai.services.ai_services import SentenceAggregator
from dailyai.services.azure_ai_services import AzureLLMService
from dailyai.services.elevenlabs_ai_service import ElevenLabsTTSService
@@ -17,29 +18,27 @@ async def main(room_url):
)
transport.mic_enabled = True
text_to_llm_queue = asyncio.Queue()
llm_to_tts_queue = asyncio.Queue()
tts = ElevenLabsTTSService(
llm_to_tts_queue, transport.get_async_send_queue(), voice_id="29vD33N1CtxCmqQRPOHJ"
)
llm = AzureLLMService(text_to_llm_queue, llm_to_tts_queue)
tts = ElevenLabsTTSService(voice_id="29vD33N1CtxCmqQRPOHJ")
llm = AzureLLMService()
messages = [{
"role": "system",
"content": "You are an LLM in a WebRTC session, and this is a 'hello world' demo. Say hello to the world."
}]
await text_to_llm_queue.put(QueueFrame(FrameType.LLM_MESSAGE, messages))
await text_to_llm_queue.put(QueueFrame(FrameType.END_STREAM, None))
llm_task = asyncio.create_task(llm.run())
tts_task = asyncio.create_task(
tts.run_to_queue(
transport.send_queue,
SentenceAggregator().run(
llm.run([QueueFrame(FrameType.LLM_MESSAGE, messages)])
)
)
)
@transport.event_handler("on_first_other_participant_joined")
async def on_first_other_participant_joined(transport):
await asyncio.gather(llm_task, tts.run())
await tts_task
# wait for the output queue to be empty, then leave the meeting
transport.output_queue.join()
transport.wait_for_send_queue_to_empty()
transport.stop()
await transport.run()

11
src/samples/theoretical-to-real/03-still-frame.py
View File

@@ -21,13 +21,14 @@ async def main(room_url):
transport.camera_width = 1024
transport.camera_height = 1024
imagegen = OpenAIImageGenService()
image_task = asyncio.create_task(imagegen.run_image_gen("a cat in the style of picasso", "1024x1024"))
imagegen = OpenAIImageGenService(image_size="1024x1024")
image_task = asyncio.create_task(
imagegen.run_to_queue(transport.send_queue, [QueueFrame(FrameType.IMAGE_DESCRIPTION, "a cat in the style of picasso")])
)
@transport.event_handler("on_participant_joined")
async def on_participant_joined(transport, participant):
(_, image_bytes) = await image_task
transport.output_queue.put(QueueFrame(FrameType.IMAGE, image_bytes))
await image_task
await transport.run()
@@ -38,6 +39,6 @@ if __name__ == "__main__":
"-u", "--url", type=str, required=True, help="URL of the Daily room to join"
)
args: argparse.Namespace = parser.parse_args()
args, unknown = parser.parse_known_args()
asyncio.run(main(args.url))

54
src/samples/theoretical-to-real/04-utterance-and-speech.py
View File

@@ -2,9 +2,11 @@ import argparse
import asyncio
import re
from dailyai.services.ai_services import SentenceAggregator
from dailyai.services.daily_transport_service import DailyTransportService
from dailyai.services.azure_ai_services import AzureLLMService, AzureTTSService
from dailyai.queue_frame import QueueFrame, FrameType
from dailyai.services.elevenlabs_ai_service import ElevenLabsTTSService
async def main(room_url:str):
global transport
@@ -22,34 +24,46 @@ async def main(room_url:str):
transport.camera_enabled = False
llm = AzureLLMService()
tts = AzureTTSService()
azure_tts = AzureTTSService()
elevenlabs_tts = ElevenLabsTTSService(voice_id="ErXwobaYiN019PkySvjV")
messages = [{"role": "system", "content": "tell the user a joke about llamas"}]
# Start a task to run the LLM to create a joke, and convert the LLM output to audio frames. This task
# will run in parallel with generating and speaking the audio for static text, so there's no delay to
# speak the LLM response.
buffer_queue = asyncio.Queue()
llm_response_task = asyncio.create_task(
elevenlabs_tts.run_to_queue(
buffer_queue,
SentenceAggregator().run(
llm.run([QueueFrame(FrameType.LLM_MESSAGE, messages)])
),
True,
)
)
@transport.event_handler("on_participant_joined")
async def on_joined(transport, participant):
if participant["id"] == transport.my_participant_id:
return
# queue two pieces of speech: one specified as a text literal,
# and one generated by an llm. We'll kick off the llm first, and let
# it generate a response while we're speaking the literal string.
#
# Note that in this case, we don't use `run_llm_async` because we're
# taking advantage of the time spent speaking the first phrase to generate
# the entire LLM response, and this happens asynchronously in a task.
llm_response_task = asyncio.create_task(llm.run_llm(
[{"role": "system", "content": "tell the user a joke about llamas"}]
))
await azure_tts.run_to_queue(
transport.send_queue,
[QueueFrame(FrameType.SENTENCE, "My friend the LLM is now going to tell a joke about llamas.")]
)
async for audio_chunk in tts.run_tts("My friend the LLM is now going to tell a joke about llamas."):
transport.output_queue.put(QueueFrame(FrameType.AUDIO, audio_chunk))
async def buffer_to_send_queue():
while True:
frame = await buffer_queue.get()
await transport.send_queue.put(frame)
buffer_queue.task_done()
if frame.frame_type == FrameType.END_STREAM:
break
llm_response = await llm_response_task
async for audio_chunk in tts.run_tts(llm_response):
transport.output_queue.put(QueueFrame(FrameType.AUDIO, audio_chunk))
await asyncio.gather(llm_response_task, buffer_to_send_queue())
# wait for the output queue to be empty, then leave the meeting
transport.output_queue.join()
transport.wait_for_send_queue_to_empty()
transport.stop()
await transport.run()
@@ -61,6 +75,6 @@ if __name__ == "__main__":
"-u", "--url", type=str, required=True, help="URL of the Daily room to join"
)
args: argparse.Namespace = parser.parse_args()
args, unknown = parser.parse_known_args()
asyncio.run(main(args.url))

7
src/samples/theoretical-to-real/05-sync-speech-and-text.py
View File

@@ -26,10 +26,9 @@ async def main(room_url):
transport.camera_height = 1024
llm = AzureLLMService()
#tts = ElevenLabsTTSService(voice_id="ErXwobaYiN019PkySvjV")
tts = ElevenLabsTTSService()
dalle = FalImageGenService()
# dalle = OpenAIImageGenService()
tts = ElevenLabsTTSService(voice_id="ErXwobaYiN019PkySvjV")
#dalle = OpenAIImageGenService(image_size="1024x1024")
# Get a complete audio chunk from the given text. Splitting this into its own
# coroutine lets us ensure proper ordering of the audio chunks on the output queue.
@@ -61,7 +60,7 @@ async def main(room_url):
tts_tasks.append(get_all_audio(sentence))
tts_tasks.insert(0, dalle.run_image_gen(image_text, "1024x1024"))
tts_tasks.insert(0, dalle.run_image_gen(image_text))
print(f"waiting for tasks to finish for {month}")
data = await asyncio.gather(