39-aws-nova-sonic.py -> 40-aws-nova-sonic.py
This commit is contained in:
vipyne
173
examples/foundational/40-aws-nova-sonic.py
Normal file
173
examples/foundational/40-aws-nova-sonic.py
Normal file
@@ -0,0 +1,173 @@
|
||||
#
|
||||
# Copyright (c) 2024–2025, Daily
|
||||
#
|
||||
# SPDX-License-Identifier: BSD 2-Clause License
|
||||
#
|
||||
|
||||
import argparse
|
||||
import os
|
||||
from datetime import datetime
|
||||
|
||||
from dotenv import load_dotenv
|
||||
from loguru import logger
|
||||
|
||||
from pipecat.adapters.schemas.function_schema import FunctionSchema
|
||||
from pipecat.adapters.schemas.tools_schema import ToolsSchema
|
||||
from pipecat.audio.vad.silero import SileroVADAnalyzer
|
||||
from pipecat.audio.vad.vad_analyzer import VADParams
|
||||
from pipecat.pipeline.pipeline import Pipeline
|
||||
from pipecat.pipeline.runner import PipelineRunner
|
||||
from pipecat.pipeline.task import PipelineParams, PipelineTask
|
||||
from pipecat.processors.aggregators.openai_llm_context import OpenAILLMContext
|
||||
from pipecat.services.aws_nova_sonic import AWSNovaSonicLLMService
|
||||
from pipecat.services.llm_service import FunctionCallParams
|
||||
from pipecat.transports.base_transport import TransportParams
|
||||
from pipecat.transports.network.small_webrtc import SmallWebRTCTransport
|
||||
from pipecat.transports.network.webrtc_connection import SmallWebRTCConnection
|
||||
|
||||
# Load environment variables
|
||||
load_dotenv(override=True)
|
||||
|
||||
|
||||
async def fetch_weather_from_api(params: FunctionCallParams):
|
||||
temperature = 75 if params.arguments["format"] == "fahrenheit" else 24
|
||||
await params.result_callback(
|
||||
{
|
||||
"conditions": "nice",
|
||||
"temperature": temperature,
|
||||
"format": params.arguments["format"],
|
||||
"timestamp": datetime.now().strftime("%Y%m%d_%H%M%S"),
|
||||
}
|
||||
)
|
||||
|
||||
|
||||
weather_function = FunctionSchema(
|
||||
name="get_current_weather",
|
||||
description="Get the current weather",
|
||||
properties={
|
||||
"location": {
|
||||
"type": "string",
|
||||
"description": "The city and state, e.g. San Francisco, CA",
|
||||
},
|
||||
"format": {
|
||||
"type": "string",
|
||||
"enum": ["celsius", "fahrenheit"],
|
||||
"description": "The temperature unit to use. Infer this from the users location.",
|
||||
},
|
||||
},
|
||||
required=["location", "format"],
|
||||
)
|
||||
|
||||
# Create tools schema
|
||||
tools = ToolsSchema(standard_tools=[weather_function])
|
||||
|
||||
|
||||
async def run_bot(webrtc_connection: SmallWebRTCConnection, _: argparse.Namespace):
|
||||
logger.info(f"Starting bot")
|
||||
|
||||
# Initialize the SmallWebRTCTransport with the connection
|
||||
transport = SmallWebRTCTransport(
|
||||
webrtc_connection=webrtc_connection,
|
||||
params=TransportParams(
|
||||
audio_in_enabled=True,
|
||||
audio_in_sample_rate=16000,
|
||||
audio_out_enabled=True,
|
||||
camera_in_enabled=False,
|
||||
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.8)),
|
||||
),
|
||||
)
|
||||
|
||||
# Specify initial system instruction.
|
||||
# HACK: note that, for now, we need to inject a special bit of text into this instruction to
|
||||
# allow the first assistant response to be programmatically triggered (which happens in the
|
||||
# on_client_connected handler, below)
|
||||
system_instruction = (
|
||||
"You are a friendly assistant. The user and you will engage in a spoken dialog exchanging "
|
||||
"the transcripts of a natural real-time conversation. Keep your responses short, generally "
|
||||
"two or three sentences for chatty scenarios. "
|
||||
f"{AWSNovaSonicLLMService.AWAIT_TRIGGER_ASSISTANT_RESPONSE_INSTRUCTION}"
|
||||
)
|
||||
|
||||
# Create the AWS Nova Sonic LLM service
|
||||
llm = AWSNovaSonicLLMService(
|
||||
secret_access_key=os.getenv("AWS_SECRET_ACCESS_KEY"),
|
||||
access_key_id=os.getenv("AWS_ACCESS_KEY_ID"),
|
||||
region=os.getenv("AWS_REGION"), # as of 2025-05-06, us-east-1 is the only supported region
|
||||
voice_id="tiffany", # matthew, tiffany, amy
|
||||
# you could choose to pass instruction here rather than via context
|
||||
# system_instruction=system_instruction
|
||||
# you could choose to pass tools here rather than via context
|
||||
# tools=tools
|
||||
)
|
||||
|
||||
# Register function for function calls
|
||||
# you can either register a single function for all function calls, or specific functions
|
||||
# llm.register_function(None, fetch_weather_from_api)
|
||||
llm.register_function("get_current_weather", fetch_weather_from_api)
|
||||
|
||||
# Set up context and context management.
|
||||
# AWSNovaSonicService will adapt OpenAI LLM context objects with standard message format to
|
||||
# what's expected by Nova Sonic.
|
||||
context = OpenAILLMContext(
|
||||
messages=[
|
||||
{"role": "system", "content": f"{system_instruction}"},
|
||||
{
|
||||
"role": "user",
|
||||
"content": "Tell me a fun fact!",
|
||||
},
|
||||
],
|
||||
tools=tools,
|
||||
)
|
||||
context_aggregator = llm.create_context_aggregator(context)
|
||||
|
||||
# Build the pipeline
|
||||
pipeline = Pipeline(
|
||||
[
|
||||
transport.input(),
|
||||
context_aggregator.user(),
|
||||
llm,
|
||||
transport.output(),
|
||||
context_aggregator.assistant(),
|
||||
]
|
||||
)
|
||||
|
||||
# Configure the pipeline task
|
||||
task = PipelineTask(
|
||||
pipeline,
|
||||
params=PipelineParams(
|
||||
allow_interruptions=True,
|
||||
enable_metrics=True,
|
||||
enable_usage_metrics=True,
|
||||
),
|
||||
)
|
||||
|
||||
# Handle client connection event
|
||||
@transport.event_handler("on_client_connected")
|
||||
async def on_client_connected(transport, client):
|
||||
logger.info(f"Client connected")
|
||||
# Kick off the conversation.
|
||||
await task.queue_frames([context_aggregator.user().get_context_frame()])
|
||||
# HACK: for now, we need this special way of triggering the first assistant response in AWS
|
||||
# Nova Sonic. Note that this trigger requires a special corresponding bit of text in the
|
||||
# system instruction. In the future, simply queueing the context frame should be sufficient.
|
||||
await llm.trigger_assistant_response()
|
||||
|
||||
# Handle client disconnection events
|
||||
@transport.event_handler("on_client_disconnected")
|
||||
async def on_client_disconnected(transport, client):
|
||||
logger.info(f"Client disconnected")
|
||||
|
||||
@transport.event_handler("on_client_closed")
|
||||
async def on_client_closed(transport, client):
|
||||
logger.info(f"Client closed connection")
|
||||
await task.cancel()
|
||||
|
||||
# Run the pipeline
|
||||
runner = PipelineRunner(handle_sigint=False)
|
||||
await runner.run(task)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
from run import main
|
||||
|
||||
main()
|
||||
Reference in New Issue
Block a user