[doc] bpo-45680: Disambiguate __getitem__ and __class_getitem__ in the data model (pythonGH-29389)

AlexWaygood · web-flow · commit 31b3a70edb12 · 2021-11-18T16:54:25.000+01:00
The documentation explaining Python's data model does not adequately explain the differences between ``__getitem__`` and ``__class_getitem__``, nor does it explain when each is called. There is an attempt at explaining ``__class_getitem__`` in the documentation for ``GenericAlias`` objects, but this does not give sufficient clarity into how the method works. Moreover, it is the wrong place for that information to be found; the explanation of ``__class_getitem__`` should be in the documentation explaining the data model. This PR has been split off from pythonGH-29335.
diff --git a/Doc/library/typing.rst b/Doc/library/typing.rst
@@ -256,6 +256,7 @@ called :class:`TypeVar`.
    def first(l: Sequence[T]) -> T:   # Generic function
        return l[0]
 
+.. _user-defined-generics:
 
 User-defined generic types
 ==========================
diff --git a/Doc/reference/datamodel.rst b/Doc/reference/datamodel.rst
@@ -2215,22 +2215,142 @@ case the instance is itself a class.
 Emulating generic types
 -----------------------
 
-One can implement the generic class syntax as specified by :pep:`484`
-(for example ``List[int]``) by defining a special method:
+When using :term:`type annotations<annotation>`, it is often useful to
+*parameterize* a :term:`generic type` using Python's square-brackets notation.
+For example, the annotation ``list[int]`` might be used to signify a
+:class:`list` in which all the elements are of type :class:`int`.
+
+.. seealso::
+
+   :pep:`484` - Type Hints
+      Introducing Python's framework for type annotations
+
+   :ref:`Generic Alias Types<types-genericalias>`
+      Documentation for objects representing parameterized generic classes
+
+   :ref:`Generics`, :ref:`user-defined generics<user-defined-generics>` and :class:`typing.Generic`
+      Documentation on how to implement generic classes that can be
+      parameterized at runtime and understood by static type-checkers.
+
+A class can *generally* only be parameterized if it defines the special
+class method ``__class_getitem__()``.
 
 .. classmethod:: object.__class_getitem__(cls, key)
 
    Return an object representing the specialization of a generic class
    by type arguments found in *key*.
 
-This method is looked up on the class object itself, and when defined in
-the class body, this method is implicitly a class method.  Note, this
-mechanism is primarily reserved for use with static type hints, other usage
-is discouraged.
+   When defined on a class, ``__class_getitem__()`` is automatically a class
+   method. As such, there is no need for it to be decorated with
+   :func:`@classmethod<classmethod>` when it is defined.
 
-.. seealso::
 
-   :pep:`560` - Core support for typing module and generic types
+The purpose of *__class_getitem__*
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The purpose of :meth:`~object.__class_getitem__` is to allow runtime
+parameterization of standard-library generic classes in order to more easily
+apply :term:`type hints<type hint>` to these classes.
+
+To implement custom generic classes that can be parameterized at runtime and
+understood by static type-checkers, users should either inherit from a standard
+library class that already implements :meth:`~object.__class_getitem__`, or
+inherit from :class:`typing.Generic`, which has its own implementation of
+``__class_getitem__()``.
+
+Custom implementations of :meth:`~object.__class_getitem__` on classes defined
+outside of the standard library may not be understood by third-party
+type-checkers such as mypy. Using ``__class_getitem__()`` on any class for
+purposes other than type hinting is discouraged.
+
+
+.. _classgetitem-versus-getitem:
+
+
+*__class_getitem__* versus *__getitem__*
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Usually, the :ref:`subscription<subscriptions>` of an object using square
+brackets will call the :meth:`~object.__getitem__` instance method defined on
+the object's class. However, if the object being subscribed is itself a class,
+the class method :meth:`~object.__class_getitem__` may be called instead.
+``__class_getitem__()`` should return a :ref:`GenericAlias<types-genericalias>`
+object if it is properly defined.
+
+Presented with the :term:`expression` ``obj[x]``, the Python interpreter
+follows something like the following process to decide whether
+:meth:`~object.__getitem__` or :meth:`~object.__class_getitem__` should be
+called::
+
+   from inspect import isclass
+
+   def subscribe(obj, x):
+       """Return the result of the expression `obj[x]`"""
+
+       class_of_obj = type(obj)
+
+       # If the class of obj defines __getitem__,
+       # call class_of_obj.__getitem__(obj, x)
+       if hasattr(class_of_obj, '__getitem__'):
+           return class_of_obj.__getitem__(obj, x)
+
+       # Else, if obj is a class and defines __class_getitem__,
+       # call obj.__class_getitem__(x)
+       elif isclass(obj) and hasattr(obj, '__class_getitem__'):
+           return obj.__class_getitem__(x)
+
+       # Else, raise an exception
+       else:
+           raise TypeError(
+               f"'{class_of_obj.__name__}' object is not subscriptable"
+           )
+
+In Python, all classes are themselves instances of other classes. The class of
+a class is known as that class's :term:`metaclass`, and most classes have the
+:class:`type` class as their metaclass. :class:`type` does not define
+:meth:`~object.__getitem__`, meaning that expressions such as ``list[int]``,
+``dict[str, float]`` and ``tuple[str, bytes]`` all result in
+:meth:`~object.__class_getitem__` being called::
+
+   >>> # list has class "type" as its metaclass, like most classes:
+   >>> type(list)
+   <class 'type'>
+   >>> type(dict) == type(list) == type(tuple) == type(str) == type(bytes)
+   True
+   >>> # "list[int]" calls "list.__class_getitem__(int)"
+   >>> list[int]
+   list[int]
+   >>> # list.__class_getitem__ returns a GenericAlias object:
+   >>> type(list[int])
+   <class 'types.GenericAlias'>
+
+However, if a class has a custom metaclass that defines
+:meth:`~object.__getitem__`, subscribing the class may result in different
+behaviour. An example of this can be found in the :mod:`enum` module::
+
+   >>> from enum import Enum
+   >>> class Menu(Enum):
+   ...     """A breakfast menu"""
+   ...     SPAM = 'spam'
+   ...     BACON = 'bacon'
+   ...
+   >>> # Enum classes have a custom metaclass:
+   >>> type(Menu)
+   <class 'enum.EnumMeta'>
+   >>> # EnumMeta defines __getitem__,
+   >>> # so __class_getitem__ is not called,
+   >>> # and the result is not a GenericAlias object:
+   >>> Menu['SPAM']
+   <Menu.SPAM: 'spam'>
+   >>> type(Menu['SPAM'])
+   <enum 'Menu'>
+
+
+.. seealso::
+   :pep:`560` - Core Support for typing module and generic types
+      Introducing :meth:`~object.__class_getitem__`, and outlining when a
+      :ref:`subscription<subscriptions>` results in ``__class_getitem__()``
+      being called instead of :meth:`~object.__getitem__`
 
 
 .. _callable-types:
@@ -2330,19 +2450,27 @@ through the object's keys; for sequences, it should iterate through the values.
 
 .. method:: object.__getitem__(self, key)
 
-   Called to implement evaluation of ``self[key]``. For sequence types, the
-   accepted keys should be integers and slice objects.  Note that the special
-   interpretation of negative indexes (if the class wishes to emulate a sequence
-   type) is up to the :meth:`__getitem__` method. If *key* is of an inappropriate
-   type, :exc:`TypeError` may be raised; if of a value outside the set of indexes
-   for the sequence (after any special interpretation of negative values),
-   :exc:`IndexError` should be raised. For mapping types, if *key* is missing (not
-   in the container), :exc:`KeyError` should be raised.
+   Called to implement evaluation of ``self[key]``. For :term:`sequence` types,
+   the accepted keys should be integers and slice objects.  Note that the
+   special interpretation of negative indexes (if the class wishes to emulate a
+   :term:`sequence` type) is up to the :meth:`__getitem__` method. If *key* is
+   of an inappropriate type, :exc:`TypeError` may be raised; if of a value
+   outside the set of indexes for the sequence (after any special
+   interpretation of negative values), :exc:`IndexError` should be raised. For
+   :term:`mapping` types, if *key* is missing (not in the container),
+   :exc:`KeyError` should be raised.
+
+   .. note::
+
+      :keyword:`for` loops expect that an :exc:`IndexError` will be raised for
+      illegal indexes to allow proper detection of the end of the sequence.
 
    .. note::
 
-      :keyword:`for` loops expect that an :exc:`IndexError` will be raised for illegal
-      indexes to allow proper detection of the end of the sequence.
+      When :ref:`subscripting<subscriptions>` a *class*, the special
+      class method :meth:`~object.__class_getitem__` may be called instead of
+      ``__getitem__()``. See :ref:`classgetitem-versus-getitem` for more
+      details.
 
 
 .. method:: object.__setitem__(self, key, value)
