cpython/Doc/library/abc.rst

315 lines
11 KiB
ReStructuredText
Raw Normal View History

:mod:`abc` --- Abstract Base Classes
====================================
.. module:: abc
:synopsis: Abstract base classes according to PEP 3119.
.. moduleauthor:: Guido van Rossum
.. sectionauthor:: Georg Brandl
.. much of the content adapted from docstrings
2011-01-24 12:28:06 -04:00
**Source code:** :source:`Lib/abc.py`
--------------
2011-06-15 12:49:20 -03:00
This module provides the infrastructure for defining :term:`abstract base
classes <abstract base class>` (ABCs) in Python, as outlined in :pep:`3119`;
see the PEP for why this was added to Python. (See also :pep:`3141` and the
:mod:`numbers` module regarding a type hierarchy for numbers based on ABCs.)
The :mod:`collections` module has some concrete classes that derive from
ABCs; these can, of course, be further derived. In addition the
:mod:`collections.abc` submodule has some ABCs that can be used to test whether
a class or instance provides a particular interface, for example, is it
hashable or a mapping.
This module provides the following classes:
.. class:: ABCMeta
Metaclass for defining Abstract Base Classes (ABCs).
Use this metaclass to create an ABC. An ABC can be subclassed directly, and
then acts as a mix-in class. You can also register unrelated concrete
classes (even built-in classes) and unrelated ABCs as "virtual subclasses" --
these and their descendants will be considered subclasses of the registering
ABC by the built-in :func:`issubclass` function, but the registering ABC
won't show up in their MRO (Method Resolution Order) nor will method
implementations defined by the registering ABC be callable (not even via
:func:`super`). [#]_
Classes created with a metaclass of :class:`ABCMeta` have the following method:
.. method:: register(subclass)
Register *subclass* as a "virtual subclass" of this ABC. For
example::
from abc import ABCMeta
class MyABC(metaclass=ABCMeta):
pass
MyABC.register(tuple)
assert issubclass(tuple, MyABC)
assert isinstance((), MyABC)
.. versionchanged:: 3.3
Returns the registered subclass, to allow usage as a class decorator.
You can also override this method in an abstract base class:
.. method:: __subclasshook__(subclass)
(Must be defined as a class method.)
Check whether *subclass* is considered a subclass of this ABC. This means
that you can customize the behavior of ``issubclass`` further without the
need to call :meth:`register` on every class you want to consider a
2007-09-04 12:45:25 -03:00
subclass of the ABC. (This class method is called from the
:meth:`__subclasscheck__` method of the ABC.)
This method should return ``True``, ``False`` or ``NotImplemented``. If
it returns ``True``, the *subclass* is considered a subclass of this ABC.
If it returns ``False``, the *subclass* is not considered a subclass of
this ABC, even if it would normally be one. If it returns
``NotImplemented``, the subclass check is continued with the usual
mechanism.
2007-09-04 12:45:25 -03:00
.. XXX explain the "usual mechanism"
2007-09-04 12:45:25 -03:00
For a demonstration of these concepts, look at this example ABC definition::
2007-09-04 12:45:25 -03:00
class Foo:
def __getitem__(self, index):
...
def __len__(self):
...
def get_iterator(self):
return iter(self)
2007-09-04 12:45:25 -03:00
class MyIterable(metaclass=ABCMeta):
2007-09-04 12:45:25 -03:00
@abstractmethod
def __iter__(self):
2007-09-04 12:45:25 -03:00
while False:
yield None
def get_iterator(self):
return self.__iter__()
@classmethod
def __subclasshook__(cls, C):
2007-09-04 12:45:25 -03:00
if cls is MyIterable:
if any("__iter__" in B.__dict__ for B in C.__mro__):
return True
return NotImplemented
2007-09-04 12:45:25 -03:00
MyIterable.register(Foo)
2007-09-04 12:45:25 -03:00
The ABC ``MyIterable`` defines the standard iterable method,
:meth:`__iter__`, as an abstract method. The implementation given here can
still be called from subclasses. The :meth:`get_iterator` method is also
part of the ``MyIterable`` abstract base class, but it does not have to be
overridden in non-abstract derived classes.
The :meth:`__subclasshook__` class method defined here says that any class
2007-09-04 12:45:25 -03:00
that has an :meth:`__iter__` method in its :attr:`__dict__` (or in that of
one of its base classes, accessed via the :attr:`__mro__` list) is
considered a ``MyIterable`` too.
2007-09-04 12:45:25 -03:00
Finally, the last line makes ``Foo`` a virtual subclass of ``MyIterable``,
even though it does not define an :meth:`__iter__` method (it uses the
2007-09-04 12:45:25 -03:00
old-style iterable protocol, defined in terms of :meth:`__len__` and
:meth:`__getitem__`). Note that this will not make ``get_iterator``
available as a method of ``Foo``, so it is provided separately.
.. class:: ABC
A helper class that has :class:`ABCMeta` as its metaclass. With this class,
an abstract base class can be created by simply deriving from :class:`ABC`,
avoiding sometimes confusing metaclass usage.
Note that the type of :class:`ABC` is still :class:`ABCMeta`, therefore
inheriting from :class:`ABC` requires the usual precautions regarding metaclass
usage, as multiple inheritance may lead to metaclass conflicts.
.. versionadded:: 3.4
The :mod:`abc` module also provides the following decorators:
2012-12-08 19:35:31 -04:00
.. decorator:: abstractmethod
A decorator indicating abstract methods.
Using this decorator requires that the class's metaclass is :class:`ABCMeta`
or is derived from it. A class that has a metaclass derived from
:class:`ABCMeta` cannot be instantiated unless all of its abstract methods
and properties are overridden. The abstract methods can be called using any
of the normal 'super' call mechanisms. :func:`abstractmethod` may be used
to declare abstract methods for properties and descriptors.
2007-09-04 12:45:25 -03:00
Dynamically adding abstract methods to a class, or attempting to modify the
abstraction status of a method or class once it is created, are not
supported. The :func:`abstractmethod` only affects subclasses derived using
regular inheritance; "virtual subclasses" registered with the ABC's
:meth:`register` method are not affected.
When :func:`abstractmethod` is applied in combination with other method
descriptors, it should be applied as the innermost decorator, as shown in
the following usage examples::
class C(metaclass=ABCMeta):
@abstractmethod
def my_abstract_method(self, ...):
...
@classmethod
@abstractmethod
def my_abstract_classmethod(cls, ...):
...
@staticmethod
@abstractmethod
def my_abstract_staticmethod(...):
...
@property
@abstractmethod
def my_abstract_property(self):
...
@my_abstract_property.setter
@abstractmethod
def my_abstract_property(self, val):
...
@abstractmethod
def _get_x(self):
...
@abstractmethod
def _set_x(self, val):
...
x = property(_get_x, _set_x)
In order to correctly interoperate with the abstract base class machinery,
the descriptor must identify itself as abstract using
:attr:`__isabstractmethod__`. In general, this attribute should be ``True``
if any of the methods used to compose the descriptor are abstract. For
example, Python's built-in property does the equivalent of::
class Descriptor:
...
@property
def __isabstractmethod__(self):
return any(getattr(f, '__isabstractmethod__', False) for
f in (self._fget, self._fset, self._fdel))
2007-09-04 12:45:25 -03:00
.. note::
Unlike Java abstract methods, these abstract
methods may have an implementation. This implementation can be
called via the :func:`super` mechanism from the class that
overrides it. This could be useful as an end-point for a
super-call in a framework that uses cooperative
multiple-inheritance.
2007-09-04 12:45:25 -03:00
2012-12-08 19:35:31 -04:00
.. decorator:: abstractclassmethod
A subclass of the built-in :func:`classmethod`, indicating an abstract
classmethod. Otherwise it is similar to :func:`abstractmethod`.
This special case is deprecated, as the :func:`classmethod` decorator
is now correctly identified as abstract when applied to an abstract
method::
class C(metaclass=ABCMeta):
@classmethod
@abstractmethod
def my_abstract_classmethod(cls, ...):
...
2010-08-17 00:37:20 -03:00
.. versionadded:: 3.2
.. deprecated:: 3.3
It is now possible to use :class:`classmethod` with
:func:`abstractmethod`, making this decorator redundant.
2010-08-17 00:37:20 -03:00
2012-12-08 19:35:31 -04:00
.. decorator:: abstractstaticmethod
A subclass of the built-in :func:`staticmethod`, indicating an abstract
staticmethod. Otherwise it is similar to :func:`abstractmethod`.
This special case is deprecated, as the :func:`staticmethod` decorator
is now correctly identified as abstract when applied to an abstract
method::
class C(metaclass=ABCMeta):
@staticmethod
@abstractmethod
def my_abstract_staticmethod(...):
...
2010-08-17 00:37:20 -03:00
.. versionadded:: 3.2
.. deprecated:: 3.3
It is now possible to use :class:`staticmethod` with
:func:`abstractmethod`, making this decorator redundant.
2010-08-17 00:37:20 -03:00
.. decorator:: abstractproperty(fget=None, fset=None, fdel=None, doc=None)
A subclass of the built-in :func:`property`, indicating an abstract
property.
Using this function requires that the class's metaclass is :class:`ABCMeta`
or is derived from it. A class that has a metaclass derived from
:class:`ABCMeta` cannot be instantiated unless all of its abstract methods
and properties are overridden. The abstract properties can be called using
any of the normal 'super' call mechanisms.
This special case is deprecated, as the :func:`property` decorator
is now correctly identified as abstract when applied to an abstract
method::
class C(metaclass=ABCMeta):
@property
@abstractmethod
def my_abstract_property(self):
...
The above example defines a read-only property; you can also define a
read-write abstract property by appropriately marking one or more of the
underlying methods as abstract::
class C(metaclass=ABCMeta):
@property
def x(self):
...
@x.setter
@abstractmethod
def x(self, val):
...
If only some components are abstract, only those components need to be
updated to create a concrete property in a subclass::
class D(C):
@C.x.setter
def x(self, val):
...
.. deprecated:: 3.3
It is now possible to use :class:`property`, :meth:`property.getter`,
:meth:`property.setter` and :meth:`property.deleter` with
:func:`abstractmethod`, making this decorator redundant.
Merged revisions 70712,70714,70764-70765,70769-70771,70773,70776-70777,70788-70789,70824,70828,70832,70836,70842,70851,70855,70857,70866-70872,70883,70885,70893-70894,70896-70897,70903,70905-70907,70915,70927,70933,70951,70960,70962-70964,70998,71001,71006,71008,71010-71011,71019,71037,71056,71094,71101-71103,71106,71119,71123,71149-71150,71203,71212,71214-71217,71221,71240 via svnmerge from svn+ssh://pythondev@svn.python.org/python/trunk ........ r70712 | benjamin.peterson | 2009-03-30 10:15:38 -0500 (Mon, 30 Mar 2009) | 1 line don't rely on the order dict repr #5605 ........ r70714 | brett.cannon | 2009-03-30 10:20:53 -0500 (Mon, 30 Mar 2009) | 1 line Add an entry to developers.txt. ........ r70764 | martin.v.loewis | 2009-03-30 17:06:33 -0500 (Mon, 30 Mar 2009) | 2 lines Add several VM developers. ........ r70765 | georg.brandl | 2009-03-30 17:09:34 -0500 (Mon, 30 Mar 2009) | 1 line #5199: make warning about vars() assignment more visible. ........ r70769 | andrew.kuchling | 2009-03-30 17:29:53 -0500 (Mon, 30 Mar 2009) | 1 line Remove comment ........ r70770 | andrew.kuchling | 2009-03-30 17:30:20 -0500 (Mon, 30 Mar 2009) | 1 line Add several items and placeholders ........ r70771 | andrew.kuchling | 2009-03-30 17:31:11 -0500 (Mon, 30 Mar 2009) | 1 line Many edits ........ r70773 | georg.brandl | 2009-03-30 17:43:00 -0500 (Mon, 30 Mar 2009) | 1 line #5039: make it clear that the impl. note refers to CPython. ........ r70776 | andrew.kuchling | 2009-03-30 18:08:24 -0500 (Mon, 30 Mar 2009) | 1 line typo fix ........ r70777 | andrew.kuchling | 2009-03-30 18:09:46 -0500 (Mon, 30 Mar 2009) | 1 line Add more items ........ r70788 | andrew.kuchling | 2009-03-30 20:21:01 -0500 (Mon, 30 Mar 2009) | 1 line Add various items ........ r70789 | georg.brandl | 2009-03-30 20:25:15 -0500 (Mon, 30 Mar 2009) | 1 line Fix a wrong struct field assignment (docstring as closure). ........ r70824 | georg.brandl | 2009-03-31 10:43:20 -0500 (Tue, 31 Mar 2009) | 1 line #5519: remove reference to Kodos, which seems dead. ........ r70828 | georg.brandl | 2009-03-31 10:50:16 -0500 (Tue, 31 Mar 2009) | 1 line #5581: fget argument of abstractproperty is optional as well. ........ r70832 | georg.brandl | 2009-03-31 11:31:11 -0500 (Tue, 31 Mar 2009) | 1 line #1386675: specify WindowsError as the exception, because it has a winerror attribute that EnvironmentError doesnt have. ........ r70836 | georg.brandl | 2009-03-31 11:50:25 -0500 (Tue, 31 Mar 2009) | 1 line #5417: replace references to undocumented functions by ones to documented functions. ........ r70842 | georg.brandl | 2009-03-31 12:13:06 -0500 (Tue, 31 Mar 2009) | 1 line #970783: document PyObject_Generic[GS]etAttr. ........ r70851 | georg.brandl | 2009-03-31 13:26:55 -0500 (Tue, 31 Mar 2009) | 1 line #837577: note cryptic return value of spawn*e on invalid env dicts. ........ r70855 | georg.brandl | 2009-03-31 13:30:37 -0500 (Tue, 31 Mar 2009) | 1 line #5245: note that PyRun_SimpleString doesnt return on SystemExit. ........ r70857 | georg.brandl | 2009-03-31 13:33:10 -0500 (Tue, 31 Mar 2009) | 1 line #5227: note that Py_Main doesnt return on SystemExit. ........ r70866 | georg.brandl | 2009-03-31 14:06:57 -0500 (Tue, 31 Mar 2009) | 1 line #4882: document named group behavior a bit better. ........ r70867 | georg.brandl | 2009-03-31 14:10:35 -0500 (Tue, 31 Mar 2009) | 1 line #1096310: document usage of sys.__std*__ a bit better. ........ r70868 | georg.brandl | 2009-03-31 14:12:17 -0500 (Tue, 31 Mar 2009) | 1 line #5190: export make_option in __all__. ........ r70869 | georg.brandl | 2009-03-31 14:14:42 -0500 (Tue, 31 Mar 2009) | 1 line Fix-up unwanted change. ........ r70870 | georg.brandl | 2009-03-31 14:26:24 -0500 (Tue, 31 Mar 2009) | 1 line #4411: document mro() and __mro__. (I hope I got it right.) ........ r70871 | georg.brandl | 2009-03-31 14:30:56 -0500 (Tue, 31 Mar 2009) | 1 line #5618: fix typo. ........ r70872 | r.david.murray | 2009-03-31 14:31:17 -0500 (Tue, 31 Mar 2009) | 3 lines Delete out-of-date and little-known README from the test directory by consensus of devs at pycon sprint. ........ r70883 | georg.brandl | 2009-03-31 15:41:08 -0500 (Tue, 31 Mar 2009) | 1 line #1674032: return value of flag from Event.wait(). OKed by Guido. ........ r70885 | tarek.ziade | 2009-03-31 15:48:31 -0500 (Tue, 31 Mar 2009) | 1 line using log.warn for sys.stderr ........ r70893 | georg.brandl | 2009-03-31 15:56:32 -0500 (Tue, 31 Mar 2009) | 1 line #1530012: move TQS section before raw strings. ........ r70894 | benjamin.peterson | 2009-03-31 16:06:30 -0500 (Tue, 31 Mar 2009) | 1 line take the usual lock precautions around _active_limbo_lock ........ r70896 | georg.brandl | 2009-03-31 16:15:33 -0500 (Tue, 31 Mar 2009) | 1 line #5598: document DocFileSuite *args argument. ........ r70897 | benjamin.peterson | 2009-03-31 16:34:42 -0500 (Tue, 31 Mar 2009) | 1 line fix Thread.ident when it is the main thread or a dummy thread #5632 ........ r70903 | georg.brandl | 2009-03-31 16:45:18 -0500 (Tue, 31 Mar 2009) | 1 line #1676135: remove trailing slashes from --prefix argument. ........ r70905 | georg.brandl | 2009-03-31 17:03:40 -0500 (Tue, 31 Mar 2009) | 1 line #5563: more documentation for bdist_msi. ........ r70906 | georg.brandl | 2009-03-31 17:11:53 -0500 (Tue, 31 Mar 2009) | 1 line #1651995: fix _convert_ref for non-ASCII characters. ........ r70907 | georg.brandl | 2009-03-31 17:18:19 -0500 (Tue, 31 Mar 2009) | 1 line #3427: document correct return type for urlopen().info(). ........ r70915 | georg.brandl | 2009-03-31 17:40:16 -0500 (Tue, 31 Mar 2009) | 1 line #5018: remove confusing paragraph. ........ r70927 | georg.brandl | 2009-03-31 18:01:27 -0500 (Tue, 31 Mar 2009) | 1 line Dont shout to users. ........ r70933 | georg.brandl | 2009-03-31 19:04:33 -0500 (Tue, 31 Mar 2009) | 2 lines Issue #5635: Fix running test_sys with tracing enabled. ........ r70951 | georg.brandl | 2009-04-01 09:02:27 -0500 (Wed, 01 Apr 2009) | 1 line Add Maksim, who worked on several issues at the sprint. ........ r70960 | jesse.noller | 2009-04-01 11:42:19 -0500 (Wed, 01 Apr 2009) | 1 line Issue 3270: document Listener address restrictions on windows ........ r70962 | brett.cannon | 2009-04-01 12:07:16 -0500 (Wed, 01 Apr 2009) | 2 lines Ron DuPlain was given commit privileges at PyCon 2009 to work on 3to2. ........ r70963 | georg.brandl | 2009-04-01 12:46:01 -0500 (Wed, 01 Apr 2009) | 1 line #5655: fix docstring oversight. ........ r70964 | brett.cannon | 2009-04-01 12:52:13 -0500 (Wed, 01 Apr 2009) | 2 lines Paul Kippes was given commit privileges to work on 3to2. ........ r70998 | georg.brandl | 2009-04-01 16:54:21 -0500 (Wed, 01 Apr 2009) | 1 line In Pdb, stop assigning values to __builtin__._ which interferes with the one commonly installed by gettext. ........ r71001 | brett.cannon | 2009-04-01 18:01:12 -0500 (Wed, 01 Apr 2009) | 3 lines Add my initials to Misc/developers.txt. Names are now sorted by number of characters in the person's name. ........ r71006 | georg.brandl | 2009-04-01 18:32:17 -0500 (Wed, 01 Apr 2009) | 1 line Cache the f_locals dict of the current frame, since every access to frame.f_locals overrides its contents with the real locals which undoes modifications made by the debugging user. ........ r71008 | andrew.kuchling | 2009-04-01 19:02:14 -0500 (Wed, 01 Apr 2009) | 1 line Typo fix ........ r71010 | benjamin.peterson | 2009-04-01 19:11:52 -0500 (Wed, 01 Apr 2009) | 1 line fix markup ........ r71011 | benjamin.peterson | 2009-04-01 19:12:47 -0500 (Wed, 01 Apr 2009) | 1 line this should be :noindex: ........ r71019 | georg.brandl | 2009-04-01 21:00:01 -0500 (Wed, 01 Apr 2009) | 1 line Fix test_doctest, missed two assignments to curframe. ........ r71037 | r.david.murray | 2009-04-01 23:34:04 -0500 (Wed, 01 Apr 2009) | 6 lines Clarify that datetime strftime does not produce leap seconds and datetime strptime does not accept it in the strftime behavior section of the datetime docs. Closes issue 2568. ........ r71056 | georg.brandl | 2009-04-02 12:43:07 -0500 (Thu, 02 Apr 2009) | 2 lines Actually the displayhook should print the repr. ........ r71094 | vinay.sajip | 2009-04-03 05:23:18 -0500 (Fri, 03 Apr 2009) | 1 line Added warning about logging use from asynchronous signal handlers. ........ r71101 | andrew.kuchling | 2009-04-03 16:43:00 -0500 (Fri, 03 Apr 2009) | 1 line Add some items ........ r71102 | andrew.kuchling | 2009-04-03 16:44:49 -0500 (Fri, 03 Apr 2009) | 1 line Fix 'the the'; grammar fix ........ r71103 | andrew.kuchling | 2009-04-03 16:45:29 -0500 (Fri, 03 Apr 2009) | 1 line Fix 'the the' duplication ........ r71106 | vinay.sajip | 2009-04-03 16:58:16 -0500 (Fri, 03 Apr 2009) | 1 line Clarified warning about logging use from asynchronous signal handlers. ........ r71119 | raymond.hettinger | 2009-04-04 00:37:47 -0500 (Sat, 04 Apr 2009) | 1 line Add helpful link. ........ r71123 | r.david.murray | 2009-04-04 01:39:56 -0500 (Sat, 04 Apr 2009) | 2 lines Fix error in description of 'oct' (issue 5678). ........ r71149 | georg.brandl | 2009-04-04 08:42:39 -0500 (Sat, 04 Apr 2009) | 1 line #5642: clarify map() compatibility to the builtin. ........ r71150 | georg.brandl | 2009-04-04 08:45:49 -0500 (Sat, 04 Apr 2009) | 1 line #5601: clarify that webbrowser is not meant for file names. ........ r71203 | benjamin.peterson | 2009-04-04 18:46:34 -0500 (Sat, 04 Apr 2009) | 1 line note how using iter* are unsafe while mutating and document iter(dict) ........ r71212 | georg.brandl | 2009-04-05 05:24:20 -0500 (Sun, 05 Apr 2009) | 1 line #1742837: expand HTTP server docs, and fix SocketServer ones to document methods as methods, not functions. ........ r71214 | georg.brandl | 2009-04-05 05:29:57 -0500 (Sun, 05 Apr 2009) | 1 line Normalize spelling of Mac OS X. ........ r71215 | georg.brandl | 2009-04-05 05:32:26 -0500 (Sun, 05 Apr 2009) | 1 line Avoid sure signs of a diseased mind. ........ r71216 | georg.brandl | 2009-04-05 05:41:02 -0500 (Sun, 05 Apr 2009) | 1 line #1718017: document the relation of os.path and the posixpath, ntpath etc. modules better. ........ r71217 | georg.brandl | 2009-04-05 05:48:47 -0500 (Sun, 05 Apr 2009) | 1 line #1726172: dont raise an unexpected IndexError if a voidresp() call has an empty response. ........ r71221 | vinay.sajip | 2009-04-05 06:06:24 -0500 (Sun, 05 Apr 2009) | 1 line Issue #5695: Moved logging.captureWarnings() call inside with statement in WarningsTest.test_warnings. ........ r71240 | georg.brandl | 2009-04-05 09:40:06 -0500 (Sun, 05 Apr 2009) | 1 line #5370: doc update about unpickling objects with custom __getattr__ etc. methods. ........
2009-04-05 16:13:16 -03:00
.. rubric:: Footnotes
.. [#] C++ programmers should note that Python's virtual base class
concept is not the same as C++'s.