ytptube/app/library/AsyncPool.py

320 lines
11 KiB
Python

import asyncio
import logging
import uuid
from datetime import UTC, datetime
class Terminator:
pass
class AsyncPool:
def __init__(
self,
num_workers: int,
worker_co,
name: str,
logger: logging.Logger,
loop: asyncio.AbstractEventLoop | None = None,
load_factor: int = 1,
max_task_time: int | None = None,
return_futures: bool = False,
raise_on_join: bool = False,
):
"""
This class will create `num_workers` asyncio tasks to work against a queue of
`num_workers * load_factor` items of back-pressure (IOW we will block after such
number of items of work is in the queue). `worker_co` will be called
against each item retrieved from the queue. If any exceptions are raised out of
worker_co, self.exceptions will be set to True.
Args:
num_workers (int): number of async tasks which will pull from the internal queue
worker_co (coroutine): async coroutine to call when an item is retrieved from the queue
name (str): name of the worker pool (used for logging)
logger (logging.Logger): logger to use
loop (asyncio.AbstractEventLoop|None): asyncio loop to use
load_factor (int): multiplier used for number of items in queue
max_task_time (int|None): maximum time allowed for each task before a CancelledError is raised in the task.
return_futures (bool): set to reture to return a future for each `push` (imposes CPU overhead)
raise_on_join (bool): raise on join if any exceptions have occurred, default is False
Returns:
AsyncPool: instance of AsyncPool
"""
self._loop = loop if loop else asyncio.get_event_loop()
self._num_workers = num_workers
self._logger = logger if logger else logging.getLogger(__name__)
self._queue = asyncio.Queue(num_workers * load_factor)
self._workers: dict[str, asyncio.Future] = {}
self._exceptions = False
self._max_task_time = max_task_time
self._return_futures = return_futures
self._raise_on_join = raise_on_join
self._name = name
self._worker_co = worker_co
self._status: dict[str, dict | None] = {}
@property
def exceptions(self):
return self._exceptions
async def _worker_loop(self, worker_id: str):
"""
The main persistent worker loop that continuously processes jobs from the shared queue.
Args:
worker_id (str): the id of the worker processing this job.
"""
while True:
item = await self._queue.get()
should_continue = await self._process_item(worker_id, item, from_queue=True)
if not should_continue:
break
async def _process_item(self, worker_id: str, item: tuple | Terminator, from_queue: bool = True) -> bool:
"""
Processes a single job item.
Args:
worker_id (str): the id of the worker processing this job.
item (tuple|Terminator): the job item (tuple of (future, args, kwargs)) or a Terminator.
from_queue (bool): indicates whether the job came from the shared queue; if True, task_done() is called.
Returns:
bool: False if the item is a Terminator (indicating termination), True otherwise.
"""
future = None
is_terminator = isinstance(item, Terminator)
try:
if is_terminator:
return False
future, args, kwargs = item
self._status[worker_id] = {
"started": self._time().isoformat(),
"data": kwargs.get("entry", {"info": {}}).info.__dict__,
}
result = await asyncio.wait_for(self._worker_co(*args, **kwargs), timeout=self._max_task_time)
if future:
future.set_result(result)
except (KeyboardInterrupt, MemoryError, SystemExit) as e:
if future:
future.set_exception(e)
self._exceptions = True
raise
except Exception as e:
self._exceptions = True
if future:
future.set_exception(e)
else:
self._logger.exception(f"Worker call failed. {e!s}")
finally:
self._status[worker_id] = None
if not is_terminator and from_queue:
self._queue.task_done()
return True
def has_open_workers(self) -> bool:
"""
Check if there are open workers.
Returns:
bool: True if there are open workers, False otherwise.
"""
return self.get_available_workers() > 0
def get_available_workers(self) -> int:
"""
Get the number of available workers.
Returns:
int: number of available workers.
"""
return sum(1 for worker_status in self._status.values() if worker_status is None)
def get_workers_status(self) -> dict[str, datetime | None]:
"""
Get the status of all workers.
Returns:
dict: dictionary of worker status.
"""
return self._status
async def __aenter__(self):
self.start()
return self
async def __aexit__(self, exc_type, exc_val, exc_tb):
await self.join()
async def push(self, *args, is_temp: bool = False, **kwargs) -> asyncio.Future:
"""
Push work to the worker_co. If is_temp is True, a temporary worker is spawned
Args:
args: position arguments to be passed to `worker_co`
is_temp (bool): flag to indicate if the job is temporary, default is False
kwargs (dict): keyword arguments to be passed to `worker_co`
Returns:
asyncio.Future: future of result.
"""
future = asyncio.futures.Future(loop=self._loop) if self._return_futures else None
if is_temp:
temp_worker_id = f"temp_worker_{uuid.uuid4()!s}"
self._logger.info(f"Creating temporary worker '{temp_worker_id}'.")
# Spawn a temporary worker that processes this one job.
task = asyncio.ensure_future(
self._process_item(temp_worker_id, (future, args, kwargs), from_queue=False), loop=self._loop
)
self._workers[temp_worker_id] = task
# Attach a callback to clean up temporary worker entries when done.
def cleanup_temp_worker(_):
self._workers.pop(temp_worker_id, None)
self._status.pop(temp_worker_id, None)
self._logger.info(f"Temporary worker '{temp_worker_id}' has terminated.")
task.add_done_callback(cleanup_temp_worker)
return future
await self._queue.put((future, args, kwargs))
self._logger.debug(f"'{self._name}' pool has received a new job. {args} {kwargs}")
return future
def start(self):
"""
Will start up worker pool and reset exception state
"""
self._exceptions = False
for worker_number in range(self._num_workers):
worker_id = f"worker_{worker_number+1}"
self._create_worker(worker_id)
async def restart(self, worker_id: str, msg: str | None = None) -> bool:
"""
Will restart the worker pool
Args:
worker_id (str): worker id to restart
msg (str|None): message to send to the worker, default is None
Returns:
bool: True if worker was restarted.
"""
if worker_id not in self._workers:
self._logger.warning(f"Worker {worker_id} does not exist.")
return False
try:
self._workers[worker_id].cancel(msg)
await self._workers[worker_id]
except asyncio.exceptions.CancelledError as e:
self._logger.warning(f"Worker {worker_id} restarted. {e!s}")
if worker_id in self._status:
self._status.pop(worker_id)
if worker_id in self._workers:
self._workers.pop(worker_id)
self._create_worker(worker_id)
return True
async def on_shutdown(self, _) -> None:
"""
Will shutdown the worker pool
"""
try:
await asyncio.wait_for(asyncio.gather(*[self.stop(worker_id) for worker_id in self._workers]), timeout=10)
except Exception:
self._logger.error(f"Exception shutting down {self._name}")
async def stop(self, worker_id: str, msg: str | None = None) -> bool:
"""
Will stop the worker
Args:
worker_id (str): worker id to stop
msg (str|None): message to send to the worker, default is None
Returns:
bool: True if worker was stopped.
"""
if worker_id not in self._workers:
self._logger.warning(f"Worker {worker_id} does not exist.")
return False
if self._workers[worker_id].cancel(msg):
try:
await self._workers[worker_id]
except asyncio.exceptions.CancelledError as e:
self._logger.warning(f"Worker {worker_id} stopped. {e!s}")
if worker_id in self._status:
self._status.pop(worker_id)
if worker_id in self._workers:
self._workers.pop(worker_id)
return True
async def join(self) -> None:
# no-op if workers aren't running
if len(self._workers) < 1:
return
self._logger.info(f"Joining {self._name}")
# The Terminators will kick each worker from being blocked against the _queue.get() and allow
# each one to exit.
for _ in range(self._num_workers):
await self._queue.put(Terminator())
try:
await asyncio.gather(*list(self._workers))
self._workers = {}
except:
self._logger.exception(f"Exception joining {self._name}")
raise
finally:
self._logger.info(f"Completed {self._name}")
if self._exceptions and self._raise_on_join:
msg = f"Exception occurred in {self._name} pool"
raise Exception(msg)
def _create_worker(self, worker_id: str) -> asyncio.Future:
if worker_id in self._workers:
self._logger.debug(f"Worker {worker_id} already exists.")
return self._workers[worker_id]
self._status[worker_id] = None
self._workers[worker_id] = asyncio.ensure_future(
coro_or_future=self._worker_loop(worker_id=worker_id), loop=self._loop
)
self._logger.debug(f"Created {worker_id}")
return self._workers[worker_id]
def _time(self) -> datetime:
return datetime.now(tz=UTC)