8.8. weakref — 弱引用

源码： Lib/weakref.py

---

weakref 模块允许Python程序员创建对象的 weak references 。

在下文中，术语 referent 表示由弱引用引用的对象。

对对象的弱引用不能保证对象存活：当对像的引用只剩弱引用时， garbage collection 可以销毁引用并将其内存重用于其他内容。但是，在实际销毁对象之前，即使没有强引用，弱引用也一直能返回该对象。

弱引用的主要用途是实现保存大对象的高速缓存或映射，但又并希望大对象仅仅因为它出现在高速缓存或映射中而保持存活。

例如，如果您有许多大型二进制图像对象，则可能希望将名称与每个对象关联起来。如果您使用Python字典将名称映射到图像，或将图像映射到名称，则图像对象将保持活动状态，因为它们在字典中显示为值或键。 weakref 模块提供的 WeakKeyDictionary 和 WeakValueDictionary 类可以替代Python字典，使用弱引用来构造映射，这些映射不会仅仅因为它们出现在映射对象中而使对象保持存活。例如，如果一个图像对象是 WeakValueDictionary 中的值，那么当对该图像对象的剩余引用是弱映射对象所持有的弱引用时，垃圾回收可以回收该对象并将其在弱映射对象中相应的条目删除。

WeakKeyDictionary 和 WeakValueDictionary 在它们的实现中使用弱引用，在弱引用上设置回调函数，当键或值被垃圾回收回收时通知弱字典。 WeakSet 实现了 set 接口，但像 WeakKeyDictionary 一样，只持有其元素的弱引用。`

finalize 提供了注册一个对象被垃圾收集时要调用的清理函数的方式。这比在原始弱引用上设置回调函数更简单，因为模块会自动确保对象被回收前终结器一直保持存活。

这些弱容器类型之一或者 finalize 就是大多数程序所需要的 - 通常不需要直接创建自己的弱引用。weakref 模块暴露了低级机制，以便于高级用途。

并非所有对象都可以被弱引用；那些可以包含类实例的对象，用 Python 编写的函数（不是用C语言编写的），实例方法，集合，frozensets，某些 文件对象 ， generator，type 对象，套接字，数组，双端队列，正则表达式模式对象和代码对象。

在 3.2 版更改: 添加了对thread.lock，threading.Lock和代码对象的支持。

几个内建类型如 list 和 dict 不直接支持弱引用，但可以通过子类化添加支持:

class Dict(dict):
    pass

obj = Dict(red=1, green=2, blue=3)   # this object is weak referenceable

其他内建类型如 tuple 和 int 即使在子类化时也不支持弱引用（这是一个实现细节，在各种Python实现中可能不同）。

Extension types can easily be made to support weak references; see Weak Reference Support.

class weakref.ref(object[, callback])

返回对 对象 的弱引用。如果原始对象仍然存活，则可以通过调用引用对象来检索原始对象；如果引用的原始对象不再存在，则调用引用对象将得到 None 。如果提供了 回调 而且值不是 None ，并且返回的弱引用对象仍然存活，则在对象即将终结时将调用回调;弱引用对象将作为回调的唯一参数传递；指示物将不再可用。

It is allowable for many weak references to be constructed for the same object. Callbacks registered for each weak reference will be called from the most recently registered callback to the oldest registered callback.

Exceptions raised by the callback will be noted on the standard error output, but cannot be propagated; they are handled in exactly the same way as exceptions raised from an object’s __del__() method.

Weak references are hashable if the object is hashable. They will maintain their hash value even after the object was deleted. If hash() is called the first time only after the object was deleted, the call will raise TypeError.

Weak references support tests for equality, but not ordering. If the referents are still alive, two references have the same equality relationship as their referents (regardless of the callback). If either referent has been deleted, the references are equal only if the reference objects are the same object.

This is a subclassable type rather than a factory function.

__callback__

This read-only attribute returns the callback currently associated to the weakref. If there is no callback or if the referent of the weakref is no longer alive then this attribute will have value None.

在 3.4 版更改: Added the __callback__ attribute.

weakref.proxy(object[, callback])

Return a proxy to object which uses a weak reference. This supports use of the proxy in most contexts instead of requiring the explicit dereferencing used with weak reference objects. The returned object will have a type of either ProxyType or CallableProxyType, depending on whether object is callable. Proxy objects are not hashable regardless of the referent; this avoids a number of problems related to their fundamentally mutable nature, and prevent their use as dictionary keys. callback is the same as the parameter of the same name to the ref() function.

weakref.getweakrefcount(object)

Return the number of weak references and proxies which refer to object.

weakref.getweakrefs(object)

Return a list of all weak reference and proxy objects which refer to object.

class weakref.WeakKeyDictionary([dict])

Mapping class that references keys weakly. Entries in the dictionary will be discarded when there is no longer a strong reference to the key. This can be used to associate additional data with an object owned by other parts of an application without adding attributes to those objects. This can be especially useful with objects that override attribute accesses.

注解

Caution: Because a WeakKeyDictionary is built on top of a Python dictionary, it must not change size when iterating over it. This can be difficult to ensure for a WeakKeyDictionary because actions performed by the program during iteration may cause items in the dictionary to vanish “by magic” (as a side effect of garbage collection).

WeakKeyDictionary objects have an additional method that exposes the internal references directly. The references are not guaranteed to be “live” at the time they are used, so the result of calling the references needs to be checked before being used. This can be used to avoid creating references that will cause the garbage collector to keep the keys around longer than needed.

WeakKeyDictionary.keyrefs()

Return an iterable of the weak references to the keys.

class weakref.WeakValueDictionary([dict])

Mapping class that references values weakly. Entries in the dictionary will be discarded when no strong reference to the value exists any more.

注解

Caution: Because a WeakValueDictionary is built on top of a Python dictionary, it must not change size when iterating over it. This can be difficult to ensure for a WeakValueDictionary because actions performed by the program during iteration may cause items in the dictionary to vanish “by magic” (as a side effect of garbage collection).

WeakValueDictionary objects have an additional method that has the same issues as the keyrefs() method of WeakKeyDictionary objects.

WeakValueDictionary.valuerefs()

Return an iterable of the weak references to the values.

class weakref.WeakSet([elements])

Set class that keeps weak references to its elements. An element will be discarded when no strong reference to it exists any more.

class weakref.WeakMethod(method)

A custom ref subclass which simulates a weak reference to a bound method (i.e., a method defined on a class and looked up on an instance). Since a bound method is ephemeral, a standard weak reference cannot keep hold of it. WeakMethod has special code to recreate the bound method until either the object or the original function dies:

>>> class C:
...     def method(self):
...         print("method called!")
...
>>> c = C()
>>> r = weakref.ref(c.method)
>>> r()
>>> r = weakref.WeakMethod(c.method)
>>> r()
<bound method C.method of <__main__.C object at 0x7fc859830220>>
>>> r()()
method called!
>>> del c
>>> gc.collect()
0
>>> r()
>>>

3.4 新版功能.

class weakref.finalize(obj, func, *args, **kwargs)

Return a callable finalizer object which will be called when obj is garbage collected. Unlike an ordinary weak reference, a finalizer will always survive until the reference object is collected, greatly simplifying lifecycle management.

A finalizer is considered alive until it is called (either explicitly or at garbage collection), and after that it is dead. Calling a live finalizer returns the result of evaluating func(*arg, **kwargs), whereas calling a dead finalizer returns None.

Exceptions raised by finalizer callbacks during garbage collection will be shown on the standard error output, but cannot be propagated. They are handled in the same way as exceptions raised from an object’s __del__() method or a weak reference’s callback.

When the program exits, each remaining live finalizer is called unless its atexit attribute has been set to false. They are called in reverse order of creation.

A finalizer will never invoke its callback during the later part of the interpreter shutdown when module globals are liable to have been replaced by None.

__call__()

If self is alive then mark it as dead and return the result of calling func(*args, **kwargs). If self is dead then return None.

detach()

If self is alive then mark it as dead and return the tuple (obj, func, args, kwargs). If self is dead then return None.

peek()

If self is alive then return the tuple (obj, func, args, kwargs). If self is dead then return None.

alive

Property which is true if the finalizer is alive, false otherwise.

atexit

A writable boolean property which by default is true. When the program exits, it calls all remaining live finalizers for which atexit is true. They are called in reverse order of creation.

注解

It is important to ensure that func, args and kwargs do not own any references to obj, either directly or indirectly, since otherwise obj will never be garbage collected. In particular, func should not be a bound method of obj.

3.4 新版功能.

weakref.ReferenceType

The type object for weak references objects.

weakref.ProxyType

The type object for proxies of objects which are not callable.

weakref.CallableProxyType

The type object for proxies of callable objects.

weakref.ProxyTypes

Sequence containing all the type objects for proxies. This can make it simpler to test if an object is a proxy without being dependent on naming both proxy types.

exception weakref.ReferenceError

Exception raised when a proxy object is used but the underlying object has been collected. This is the same as the standard ReferenceError exception.

参见

PEP 205 - Weak References

The proposal and rationale for this feature, including links to earlier implementations and information about similar features in other languages.

8.8.1. 弱引用对象

Weak reference objects have no methods and no attributes besides ref.__callback__. A weak reference object allows the referent to be obtained, if it still exists, by calling it:

>>> import weakref
>>> class Object:
...     pass
...
>>> o = Object()
>>> r = weakref.ref(o)
>>> o2 = r()
>>> o is o2
True

If the referent no longer exists, calling the reference object returns None:

>>> del o, o2
>>> print(r())
None

Testing that a weak reference object is still live should be done using the expression ref() is not None. Normally, application code that needs to use a reference object should follow this pattern:

# r is a weak reference object
o = r()
if o is None:
    # referent has been garbage collected
    print("Object has been deallocated; can't frobnicate.")
else:
    print("Object is still live!")
    o.do_something_useful()

Using a separate test for “liveness” creates race conditions in threaded applications; another thread can cause a weak reference to become invalidated before the weak reference is called; the idiom shown above is safe in threaded applications as well as single-threaded applications.

Specialized versions of ref objects can be created through subclassing. This is used in the implementation of the WeakValueDictionary to reduce the memory overhead for each entry in the mapping. This may be most useful to associate additional information with a reference, but could also be used to insert additional processing on calls to retrieve the referent.

This example shows how a subclass of ref can be used to store additional information about an object and affect the value that’s returned when the referent is accessed:

import weakref

class ExtendedRef(weakref.ref):
    def __init__(self, ob, callback=None, **annotations):
        super(ExtendedRef, self).__init__(ob, callback)
        self.__counter = 0
        for k, v in annotations.items():
            setattr(self, k, v)

    def __call__(self):
        """Return a pair containing the referent and the number of
        times the reference has been called.
        """
        ob = super(ExtendedRef, self).__call__()
        if ob is not None:
            self.__counter += 1
            ob = (ob, self.__counter)
        return ob

8.8.2. 示例

This simple example shows how an application can use object IDs to retrieve objects that it has seen before. The IDs of the objects can then be used in other data structures without forcing the objects to remain alive, but the objects can still be retrieved by ID if they do.

import weakref

_id2obj_dict = weakref.WeakValueDictionary()

def remember(obj):
    oid = id(obj)
    _id2obj_dict[oid] = obj
    return oid

def id2obj(oid):
    return _id2obj_dict[oid]

8.8.3. Finalizer Objects

The main benefit of using finalize is that it makes it simple to register a callback without needing to preserve the returned finalizer object. For instance

>>> import weakref
>>> class Object:
...     pass
...
>>> kenny = Object()
>>> weakref.finalize(kenny, print, "You killed Kenny!")  #doctest:+ELLIPSIS
<finalize object at ...; for 'Object' at ...>
>>> del kenny
You killed Kenny!

The finalizer can be called directly as well. However the finalizer will invoke the callback at most once.

>>> def callback(x, y, z):
...     print("CALLBACK")
...     return x + y + z
...
>>> obj = Object()
>>> f = weakref.finalize(obj, callback, 1, 2, z=3)
>>> assert f.alive
>>> assert f() == 6
CALLBACK
>>> assert not f.alive
>>> f()                     # callback not called because finalizer dead
>>> del obj                 # callback not called because finalizer dead

You can unregister a finalizer using its detach() method. This kills the finalizer and returns the arguments passed to the constructor when it was created.

>>> obj = Object()
>>> f = weakref.finalize(obj, callback, 1, 2, z=3)
>>> f.detach()                                           #doctest:+ELLIPSIS
(<__main__.Object object ...>, <function callback ...>, (1, 2), {'z': 3})
>>> newobj, func, args, kwargs = _
>>> assert not f.alive
>>> assert newobj is obj
>>> assert func(*args, **kwargs) == 6
CALLBACK

Unless you set the atexit attribute to False, a finalizer will be called when the program exits if it is still alive. For instance

>>> obj = Object()
>>> weakref.finalize(obj, print, "obj dead or exiting")  #doctest:+ELLIPSIS
<finalize object at ...; for 'Object' at ...>
>>> exit()                                               #doctest:+SKIP
obj dead or exiting

8.8.4. Comparing finalizers with __del__() methods

Suppose we want to create a class whose instances represent temporary directories. The directories should be deleted with their contents when the first of the following events occurs:

• the object is garbage collected,
• the object’s remove() method is called, or
• the program exits.

We might try to implement the class using a __del__() method as follows:

class TempDir:
    def __init__(self):
        self.name = tempfile.mkdtemp()

    def remove(self):
        if self.name is not None:
            shutil.rmtree(self.name)
            self.name = None

    @property
    def removed(self):
        return self.name is None

    def __del__(self):
        self.remove()

Starting with Python 3.4, __del__() methods no longer prevent reference cycles from being garbage collected, and module globals are no longer forced to None during interpreter shutdown. So this code should work without any issues on CPython.

However, handling of __del__() methods is notoriously implementation specific, since it depends on internal details of the interpreter’s garbage collector implementation.

A more robust alternative can be to define a finalizer which only references the specific functions and objects that it needs, rather than having access to the full state of the object:

class TempDir:
    def __init__(self):
        self.name = tempfile.mkdtemp()
        self._finalizer = weakref.finalize(self, shutil.rmtree, self.name)

    def remove(self):
        self._finalizer()

    @property
    def removed(self):
        return not self._finalizer.alive

Defined like this, our finalizer only receives a reference to the details it needs to clean up the directory appropriately. If the object never gets garbage collected the finalizer will still be called at exit.

The other advantage of weakref based finalizers is that they can be used to register finalizers for classes where the definition is controlled by a third party, such as running code when a module is unloaded:

import weakref, sys
def unloading_module():
    # implicit reference to the module globals from the function body
weakref.finalize(sys.modules[__name__], unloading_module)

注解

If you create a finalizer object in a daemonic thread just as the program exits then there is the possibility that the finalizer does not get called at exit. However, in a daemonic thread atexit.register(), try: ... finally: ... and with: ... do not guarantee that cleanup occurs either.