同步化原始物件 (Synchronization Primitives)¶
asyncio 的同步化原始物件被設計成和那些 threading
模組 (module) 中的同名物件相似,但有兩個重要的限制條件:
asyncio 原始物件並不支援執行緒安全 (thread-safe),因此他們不可被用於 OS 執行緒同步化(請改用
threading
);這些同步化原始物件的方法 (method) 並不接受 timeout 引數;要達成有超時 (timeout) 設定的操作請改用
asyncio.wait_for()
函式。
asyncio 有以下基礎同步化原始物件:
Lock¶
- class asyncio.Lock¶
實作了一個給 asyncio 任務 (task) 用的互斥鎖 (mutex lock)。不支援執行緒安全。
一個 asyncio 的鎖可以用來確保一個共享資源的存取權被獨佔。
使用 Lock 的推薦方式是透過
async with
陳述式:lock = asyncio.Lock() # ... later async with lock: # access shared state
這等價於:
lock = asyncio.Lock() # ... later await lock.acquire() try: # access shared state finally: lock.release()
在 3.10 版的變更: 移除
loop
參數。- coroutine acquire()¶
獲得鎖。
此方法會持續等待直到鎖的狀態成為 unlocked,並將其設置為 locked 和回傳
True
。當多於一個的協程 (coroutine) 在
acquire()
中等待解鎖而被阻塞,最終只會有其中的一個被處理。鎖的獲取方式是公平的:被處理的協程會是最早開始等待解鎖的那一個。
- release()¶
釋放鎖。
如果鎖的狀態為 locked 則將其重置為 unlocked 並回傳。
如果鎖的狀態為 unlocked 則
RuntimeError
會被引發。
- locked()¶
如果鎖的狀態為 locked 則回傳
True
。
Event¶
- class asyncio.Event¶
一個事件 (event) 物件。不支援執行緒安全。
一個 asyncio 事件可以被用於通知多個有發生某些事件於其中的 asyncio 任務。
一個 Event 物件會管理一個內部旗標 (flag),它可以透過
set()
方法來被設為 true 並透過clear()
方法來重置為 false。wait()
方法會被阻塞 (block) 直到該旗標被設為 true。該旗標初始設置為 false。在 3.10 版的變更: 移除
loop
參數。範例:
async def waiter(event): print('waiting for it ...') await event.wait() print('... got it!') async def main(): # Create an Event object. event = asyncio.Event() # Spawn a Task to wait until 'event' is set. waiter_task = asyncio.create_task(waiter(event)) # Sleep for 1 second and set the event. await asyncio.sleep(1) event.set() # Wait until the waiter task is finished. await waiter_task asyncio.run(main())
- set()¶
設置事件。
所有正在等待事件被設置的任務會立即被喚醒。
- is_set()¶
如果事件有被設置則回傳
True
。
Condition¶
- class asyncio.Condition(lock=None)¶
一個條件 (codition) 物件。不支援執行緒安全。
一個 asyncio 條件原始物件可以被任務用來等待某事件發生,並獲得一個共享資源的獨佔存取權。
本質上,一個 Condition 物件會結合
Event
和Lock
的功能。多個 Condition 物件共享一個 Lock 是有可能發生的,這能夠協調關注同一共享資源的不同狀態以獲取其獨佔存取權的多個任務。可選的 lock 引數必須是一個
Lock
物件或者為None
。如為後者則一個新的 Lock 物件會被自動建立。在 3.10 版的變更: 移除
loop
參數。使用 Condition 的推薦方式是透過
async with
陳述式:cond = asyncio.Condition() # ... later async with cond: await cond.wait()
這等價於:
cond = asyncio.Condition() # ... later await cond.acquire() try: await cond.wait() finally: cond.release()
- coroutine acquire()¶
獲取底層的鎖。
此方法會持續等待直到底層的鎖為 unlocked,並將其設為 locked 並回傳
True
。
- notify(n=1)¶
Wake up n tasks (1 by default) waiting on this condition. If fewer than n tasks are waiting they are all awakened.
在此方法被呼叫前必須先獲得鎖,並在之後立刻將其釋放。如果呼叫於一個 unlocked 的鎖則
RuntimeError
錯誤會被引發。
- locked()¶
如果已獲取底層的鎖則回傳
True
。
- notify_all()¶
喚醒所有正在等待此條件的任務。
這個方法的行為就像
notify()
,但會喚醒所有正在等待的任務。在此方法被呼叫前必須先獲得鎖,並在之後立刻將其釋放。如果呼叫於一個 unlocked 的鎖則
RuntimeError
錯誤會被引發。
- release()¶
釋放底層的鎖。
當調用於一個未被解開的鎖之上時,會引發一個
RuntimeError
。
- coroutine wait()¶
持續等待直到被通知 (notify)。
當此方法被呼叫時,如果呼叫它的任務還沒有獲取鎖的話,
RuntimeError
會被引發。此方法會釋放底層的鎖,然後持續阻塞直到被
notify()
或notify_all()
的呼叫所喚醒。一但被喚醒,Condition 會重新獲取該鎖且此方法會回傳True
。Note that a task may return from this call spuriously, which is why the caller should always re-check the state and be prepared to
wait()
again. For this reason, you may prefer to usewait_for()
instead.
Semaphore¶
- class asyncio.Semaphore(value=1)¶
一個旗號 (semaphore) 物件。不支援執行緒安全。
一個旗號物件會管理一個內部計數器,會在每次呼叫
acquire()
時減少一、每次呼叫release()
時增加一。此計數器永遠不會少於零;當acquire()
發現它是零時,它會持續阻塞並等待某任務呼叫release()
。可選的 value 引數給定了內部計數器的初始值(預設為
1
)。如給定的值少於0
則ValueError
會被引發。在 3.10 版的變更: 移除
loop
參數。使用 Semaphore 的推薦方式是透過
async with
陳述式:sem = asyncio.Semaphore(10) # ... later async with sem: # work with shared resource
這等價於:
sem = asyncio.Semaphore(10) # ... later await sem.acquire() try: # work with shared resource finally: sem.release()
- locked()¶
如果旗號無法立即被取得則回傳
True
。
- release()¶
釋放一個旗號,並為其內部的計數器數值增加一。可以把一個正在等待獲取旗號的任務叫醒。
和
BoundedSemaphore
不同,Semaphore
允許release()
的呼叫次數多於acquire()
。
BoundedSemaphore¶
- class asyncio.BoundedSemaphore(value=1)¶
一個有界的旗號物件。不支援執行緒安全。
Bounded Semaphore 是
Semaphore
的另一版本,如果其內部的計數器數值增加至大於初始 value 值的話,ValueError
會在release()
時被引發。在 3.10 版的變更: 移除
loop
參數。
Barrier¶
- class asyncio.Barrier(parties)¶
一個屏障 (barrier) 物件。不支援執行緒安全。
A barrier is a simple synchronization primitive that allows to block until parties number of tasks are waiting on it. Tasks can wait on the
wait()
method and would be blocked until the specified number of tasks end up waiting onwait()
. At that point all of the waiting tasks would unblock simultaneously.async with
can be used as an alternative to awaiting onwait()
.The barrier can be reused any number of times.
範例:
async def example_barrier(): # barrier with 3 parties b = asyncio.Barrier(3) # create 2 new waiting tasks asyncio.create_task(b.wait()) asyncio.create_task(b.wait()) await asyncio.sleep(0) print(b) # The third .wait() call passes the barrier await b.wait() print(b) print("barrier passed") await asyncio.sleep(0) print(b) asyncio.run(example_barrier())
Result of this example is:
<asyncio.locks.Barrier object at 0x... [filling, waiters:2/3]> <asyncio.locks.Barrier object at 0x... [draining, waiters:0/3]> barrier passed <asyncio.locks.Barrier object at 0x... [filling, waiters:0/3]>
在 3.11 版被加入.
- coroutine wait()¶
Pass the barrier. When all the tasks party to the barrier have called this function, they are all unblocked simultaneously.
When a waiting or blocked task in the barrier is cancelled, this task exits the barrier which stays in the same state. If the state of the barrier is "filling", the number of waiting task decreases by 1.
The return value is an integer in the range of 0 to
parties-1
, different for each task. This can be used to select a task to do some special housekeeping, e.g.:... async with barrier as position: if position == 0: # Only one task prints this print('End of *draining phase*')
This method may raise a
BrokenBarrierError
exception if the barrier is broken or reset while a task is waiting. It could raise aCancelledError
if a task is cancelled.
- coroutine reset()¶
Return the barrier to the default, empty state. Any tasks waiting on it will receive the
BrokenBarrierError
exception.If a barrier is broken it may be better to just leave it and create a new one.
- coroutine abort()¶
Put the barrier into a broken state. This causes any active or future calls to
wait()
to fail with theBrokenBarrierError
. Use this for example if one of the tasks needs to abort, to avoid infinite waiting tasks.
- parties¶
The number of tasks required to pass the barrier.
- n_waiting¶
The number of tasks currently waiting in the barrier while filling.
- broken¶
A boolean that is
True
if the barrier is in the broken state.
- exception asyncio.BrokenBarrierError¶
This exception, a subclass of
RuntimeError
, is raised when theBarrier
object is reset or broken.
在 3.9 版的變更: 透過 await lock
或 yield from lock
和/或 with
陳述式 (with await lock
, with (yield from lock)
) 來獲取鎖的方式已被移除。請改用 async with lock
。