Files
pipecat/examples/multi-worker/distributed-handoff/pgmq-handoff/llm.py
Aleix Conchillo Flaqué b03247f360 Rename BaseTask → BaseWorker and reserve "task" for asyncio
Replaces every "task" identifier that referred to the BaseTask
abstraction with "worker". Asyncio task plumbing (asyncio.Task,
BaseTaskManager, TaskManager, create_task, cancel_task, etc.) stays
untouched. Highlights:

- Classes: BaseTask → BaseWorker, PipelineTask → PipelineWorker,
  LLMTask → LLMWorker, LLMContextTask → LLMContextWorker, TaskBus →
  WorkerBus, TaskRegistry → WorkerRegistry, TaskActivationArgs →
  WorkerActivationArgs, TaskReadyData → WorkerReadyData,
  TaskRegistryEntry → WorkerRegistryEntry, TaskObserver →
  WorkerObserver, all Bus*TaskMessage → Bus*WorkerMessage,
  BusAddTaskMessage.task field → worker, BusWorkerRegistryMessage.tasks
  field → workers.
- Methods/decorators: activate_task → activate_worker, deactivate_task
  → deactivate_worker, add_task → add_worker, watch_task →
  watch_worker, @task_ready → @worker_ready, setup_pipeline_task hook
  → setup_pipeline_worker.
- Params/fields: FrameProcessorSetup.pipeline_task and
  FunctionCallParams.pipeline_task → pipeline_worker. Parameter names
  like task_name → worker_name; spawn/run accept worker:.
- Files: pipeline/base_task.py → base_worker.py, pipeline/task.py →
  worker.py (plus a re-export shim at pipeline/task.py),
  task_observer.py → worker_observer.py, task_ready_decorator.py →
  worker_ready_decorator.py, pipecat.tasks → pipecat.workers,
  llm_task.py → llm_worker.py, llm_context_task.py →
  llm_context_worker.py, examples/multi-task → examples/multi-worker.

Back-compat:
- PipelineTask kept as a deprecated subclass of PipelineWorker that
  warns on construction.
- pipecat.pipeline.task re-exports PipelineWorker/PipelineTask/etc. so
  existing user imports keep working.
- FrameProcessor.pipeline_task kept as a deprecated property that
  forwards to pipeline_worker.

Local variables in examples that hold a worker (task = PipelineTask(...))
are renamed to worker = PipelineWorker(...). Asyncio-task locals
(runner_task, etc.) are preserved.
2026-05-21 19:07:13 -07:00

183 lines
6.6 KiB
Python

#
# Copyright (c) 2026, Daily
#
# SPDX-License-Identifier: BSD 2-Clause License
#
"""LLM worker — run on Machine B (or locally alongside ``main.py``).
A standalone process that runs one LLM worker (greeter or support)
attached to the same PGMQ-backed `WorkerBus` as the main worker.
Multiple instances can run on different machines as long as they
share a Postgres database with the PGMQ extension enabled.
Usage::
python llm.py greeter --database-url postgresql://...
python llm.py support --database-url postgresql://...
Requirements:
- OPENAI_API_KEY
- DATABASE_URL (or ``--database-url``)
"""
import argparse
import asyncio
import os
from urllib.parse import unquote, urlparse
from dotenv import load_dotenv
from loguru import logger
from pgmq.async_queue import PGMQueue
from pipecat.bus.network.pgmq import PgmqBus
from pipecat.pipeline.runner import PipelineRunner
from pipecat.services.llm_service import FunctionCallParams
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.workers.llm import LLMWorker, LLMWorkerActivationArgs, tool
load_dotenv(override=True)
WORKER_CONFIG = {
"greeter": {
"system_instruction": (
"You are a friendly greeter for Acme Corp. The available products "
"are: the Acme Rocket Boots, the Acme Invisible Paint, and the Acme "
"Tornado Kit. Ask which one they'd like to learn more about. "
"When the user picks a product or asks a question about one, "
"immediately call the transfer_to_agent tool with agent 'support'. "
"Do not answer product questions yourself. If the user says goodbye, "
"call the end_conversation tool. Do not mention transferring — just do it "
"seamlessly. Keep responses brief — this is a voice conversation."
),
"watch": ["support"],
},
"support": {
"system_instruction": (
"You are a support agent for Acme Corp. You know about three "
"products: Acme Rocket Boots (jet-powered boots, $299, run up "
"to 60 mph), Acme Invisible Paint (makes anything invisible for "
"24 hours, $49 per can), and Acme Tornado Kit (portable tornado "
"generator, $199, batteries included). Answer the user's questions "
"about these products. If the user wants to browse other products "
"or start over, call the transfer_to_agent tool with agent "
"'greeter'. If the user says goodbye, call the end_conversation "
"tool. Do not mention transferring — just do it seamlessly. "
"Keep responses brief — this is a voice conversation."
),
"watch": ["greeter"],
},
}
def pgmq_from_url(database_url: str, *, pool_size: int = 4) -> PGMQueue:
"""Build a `PGMQueue` from a Postgres DSN string."""
parsed = urlparse(database_url)
if parsed.scheme not in ("postgres", "postgresql"):
raise ValueError(f"Unsupported scheme '{parsed.scheme}' for database URL")
return PGMQueue(
host=parsed.hostname or "localhost",
port=str(parsed.port or 5432),
database=(parsed.path or "/postgres").lstrip("/") or "postgres",
username=unquote(parsed.username or "postgres"),
password=unquote(parsed.password or ""),
pool_size=pool_size,
)
class AcmeLLMTask(LLMWorker):
"""LLM worker for Acme Corp with transfer and end tools."""
def __init__(self, name: str, *, system_instruction: str, watch: list[str]):
"""Initialize the AcmeLLMTask.
Args:
name: Unique worker name (``"greeter"`` or ``"support"``).
system_instruction: System prompt for this LLM role.
watch: Sibling worker names this worker will watch via the
registry so it knows when they become available for
handoff.
"""
llm = OpenAILLMService(
name=f"{name}::OpenAILLMService",
api_key=os.environ["OPENAI_API_KEY"],
settings=OpenAILLMService.Settings(system_instruction=system_instruction),
)
super().__init__(name, llm=llm, bridged=())
self._watch = watch
async def start(self) -> None:
"""Register watches for sibling workers once ready."""
await super().start()
for worker_name in self._watch:
await self.watch_worker(worker_name)
@tool(cancel_on_interruption=False)
async def transfer_to_agent(self, params: FunctionCallParams, agent: str, reason: str):
"""Transfer the user to another agent.
Args:
agent (str): The agent to transfer to (e.g. 'greeter', 'support').
reason (str): Why the user is being transferred.
"""
logger.info(f"Task '{self.name}': transferring to '{agent}' ({reason})")
await self.activate_worker(
agent,
args=LLMWorkerActivationArgs(messages=[{"role": "developer", "content": reason}]),
deactivate_self=True,
result_callback=params.result_callback,
)
@tool
async def end_conversation(self, params: FunctionCallParams, reason: str):
"""End the conversation when the user says goodbye.
Args:
reason (str): Why the conversation is ending.
"""
logger.info(f"Task '{self.name}': ending conversation ({reason})")
await self.end(
reason=reason,
messages=[{"role": "developer", "content": reason}],
result_callback=params.result_callback,
)
async def main_async() -> None:
parser = argparse.ArgumentParser(description="LLM worker (greeter or support)")
parser.add_argument("worker", choices=list(WORKER_CONFIG), help="Which worker to run")
parser.add_argument(
"--database-url",
default=os.getenv("DATABASE_URL"),
help="PostgreSQL DSN (or set DATABASE_URL env var)",
)
parser.add_argument(
"--channel",
default=os.getenv("PGMQ_CHANNEL", "pipecat_acme"),
help="PGMQ channel prefix",
)
args = parser.parse_args()
if not args.database_url:
parser.error("--database-url is required (or set DATABASE_URL env var)")
pgmq = pgmq_from_url(args.database_url)
await pgmq.init()
bus = PgmqBus(pgmq=pgmq, channel=args.channel)
config = WORKER_CONFIG[args.worker]
worker = AcmeLLMTask(
args.worker,
system_instruction=config["system_instruction"],
watch=config["watch"],
)
runner = PipelineRunner(bus=bus, handle_sigint=True)
logger.info(f"Starting {args.worker} worker, waiting for activation...")
await runner.run(worker)
if __name__ == "__main__":
asyncio.run(main_async())