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

Don't create aiohttp sessions inside services

Browse Source
This commit is contained in:
Moishe Lettvin
2024-01-26 12:30:37 -05:00
parent fcceb32bd7
commit e81f247845
13 changed files with 401 additions and 408 deletions
Download Patch File Download Diff File Expand all files Collapse all files

89
src/samples/foundational/04-utterance-and-speech.py
View File

@@ -2,6 +2,8 @@ import argparse
import asyncio
import re
import aiohttp
from dailyai.services.daily_transport_service import DailyTransportService
from dailyai.services.azure_ai_services import AzureLLMService, AzureTTSService
from dailyai.queue_frame import EndStreamQueueFrame, LLMMessagesQueueFrame
@@ -9,58 +11,55 @@ from dailyai.services.elevenlabs_ai_service import ElevenLabsTTSService
async def main(room_url: str):
global transport
global llm
global tts
transport = DailyTransportService(
room_url,
None,
"Say Two Things Bot",
1,
)
transport.mic_enabled = True
transport.mic_sample_rate = 16000
transport.camera_enabled = False
llm = AzureLLMService()
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,
llm.run([LLMMessagesQueueFrame(messages)]),
True,
async with aiohttp.ClientSession() as session:
transport = DailyTransportService(
room_url,
None,
"Say Two Things Bot",
1,
)
)
transport.mic_enabled = True
transport.mic_sample_rate = 16000
transport.camera_enabled = False
@transport.event_handler("on_participant_joined")
async def on_joined(transport, participant):
if participant["id"] == transport.my_participant_id:
return
llm = AzureLLMService()
azure_tts = AzureTTSService()
elevenlabs_tts = ElevenLabsTTSService(session, voice_id="ErXwobaYiN019PkySvjV")
await azure_tts.say("My friend the LLM is now going to tell a joke about llamas.", transport.send_queue)
messages = [{"role": "system", "content": "tell the user a joke about llamas"}]
async def buffer_to_send_queue():
while True:
frame = await buffer_queue.get()
await transport.send_queue.put(frame)
buffer_queue.task_done()
if isinstance(frame, EndStreamQueueFrame):
break
# 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,
llm.run([LLMMessagesQueueFrame(messages)]),
True,
)
)
await asyncio.gather(llm_response_task, buffer_to_send_queue())
@transport.event_handler("on_participant_joined")
async def on_joined(transport, participant):
if participant["id"] == transport.my_participant_id:
return
await transport.stop_when_done()
await azure_tts.say("My friend the LLM is now going to tell a joke about llamas.", transport.send_queue)
await transport.run()
async def buffer_to_send_queue():
while True:
frame = await buffer_queue.get()
await transport.send_queue.put(frame)
buffer_queue.task_done()
if isinstance(frame, EndStreamQueueFrame):
break
await asyncio.gather(llm_response_task, buffer_to_send_queue())
await transport.stop_when_done()
await transport.run()
if __name__ == "__main__":