diff options
Diffstat (limited to 'devdocs/python~3.12/reference%2Fdatamodel.html')
| -rw-r--r-- | devdocs/python~3.12/reference%2Fdatamodel.html | 810 |
1 files changed, 810 insertions, 0 deletions
diff --git a/devdocs/python~3.12/reference%2Fdatamodel.html b/devdocs/python~3.12/reference%2Fdatamodel.html new file mode 100644 index 00000000..85356260 --- /dev/null +++ b/devdocs/python~3.12/reference%2Fdatamodel.html @@ -0,0 +1,810 @@ + <span id="datamodel"></span><h1> Data model</h1> <section id="objects-values-and-types"> <span id="objects"></span><h2> +<span class="section-number">3.1. </span>Objects, values and types</h2> <p id="index-0"><em class="dfn">Objects</em> are Python’s abstraction for data. All data in a Python program is represented by objects or by relations between objects. (In a sense, and in conformance to Von Neumann’s model of a “stored program computer”, code is also represented by objects.)</p> <span class="target" id="index-1"></span><p>Every object has an identity, a type and a value. An object’s <em>identity</em> never changes once it has been created; you may think of it as the object’s address in memory. The ‘<a class="reference internal" href="expressions#is"><code>is</code></a>’ operator compares the identity of two objects; the <a class="reference internal" href="../library/functions#id" title="id"><code>id()</code></a> function returns an integer representing its identity.</p> <div class="impl-detail compound"> <p><strong>CPython implementation detail:</strong> For CPython, <code>id(x)</code> is the memory address where <code>x</code> is stored.</p> </div> <p>An object’s type determines the operations that the object supports (e.g., “does it have a length?”) and also defines the possible values for objects of that type. The <a class="reference internal" href="../library/functions#type" title="type"><code>type()</code></a> function returns an object’s type (which is an object itself). Like its identity, an object’s <em class="dfn">type</em> is also unchangeable. <a class="footnote-reference brackets" href="#id17" id="id1">1</a></p> <p>The <em>value</em> of some objects can change. Objects whose value can change are said to be <em>mutable</em>; objects whose value is unchangeable once they are created are called <em>immutable</em>. (The value of an immutable container object that contains a reference to a mutable object can change when the latter’s value is changed; however the container is still considered immutable, because the collection of objects it contains cannot be changed. So, immutability is not strictly the same as having an unchangeable value, it is more subtle.) An object’s mutability is determined by its type; for instance, numbers, strings and tuples are immutable, while dictionaries and lists are mutable.</p> <p id="index-2">Objects are never explicitly destroyed; however, when they become unreachable they may be garbage-collected. An implementation is allowed to postpone garbage collection or omit it altogether — it is a matter of implementation quality how garbage collection is implemented, as long as no objects are collected that are still reachable.</p> <div class="impl-detail compound"> <p><strong>CPython implementation detail:</strong> CPython currently uses a reference-counting scheme with (optional) delayed detection of cyclically linked garbage, which collects most objects as soon as they become unreachable, but is not guaranteed to collect garbage containing circular references. See the documentation of the <a class="reference internal" href="../library/gc#module-gc" title="gc: Interface to the cycle-detecting garbage collector."><code>gc</code></a> module for information on controlling the collection of cyclic garbage. Other implementations act differently and CPython may change. Do not depend on immediate finalization of objects when they become unreachable (so you should always close files explicitly).</p> </div> <p>Note that the use of the implementation’s tracing or debugging facilities may keep objects alive that would normally be collectable. Also note that catching an exception with a ‘<a class="reference internal" href="compound_stmts#try"><code>try</code></a>…<a class="reference internal" href="compound_stmts#except"><code>except</code></a>’ statement may keep objects alive.</p> <p>Some objects contain references to “external” resources such as open files or windows. It is understood that these resources are freed when the object is garbage-collected, but since garbage collection is not guaranteed to happen, such objects also provide an explicit way to release the external resource, usually a <code>close()</code> method. Programs are strongly recommended to explicitly close such objects. The ‘<a class="reference internal" href="compound_stmts#try"><code>try</code></a>…<a class="reference internal" href="compound_stmts#finally"><code>finally</code></a>’ statement and the ‘<a class="reference internal" href="compound_stmts#with"><code>with</code></a>’ statement provide convenient ways to do this.</p> <p id="index-3">Some objects contain references to other objects; these are called <em>containers</em>. Examples of containers are tuples, lists and dictionaries. The references are part of a container’s value. In most cases, when we talk about the value of a container, we imply the values, not the identities of the contained objects; however, when we talk about the mutability of a container, only the identities of the immediately contained objects are implied. So, if an immutable container (like a tuple) contains a reference to a mutable object, its value changes if that mutable object is changed.</p> <p>Types affect almost all aspects of object behavior. Even the importance of object identity is affected in some sense: for immutable types, operations that compute new values may actually return a reference to any existing object with the same type and value, while for mutable objects this is not allowed. E.g., after <code>a = 1; b = 1</code>, <code>a</code> and <code>b</code> may or may not refer to the same object with the value one, depending on the implementation, but after <code>c = []; d = +[]</code>, <code>c</code> and <code>d</code> are guaranteed to refer to two different, unique, newly created empty lists. (Note that <code>c = d = []</code> assigns the same object to both <code>c</code> and <code>d</code>.)</p> </section> <section id="the-standard-type-hierarchy"> <span id="types"></span><h2> +<span class="section-number">3.2. </span>The standard type hierarchy</h2> <p id="index-4">Below is a list of the types that are built into Python. Extension modules (written in C, Java, or other languages, depending on the implementation) can define additional types. Future versions of Python may add types to the type hierarchy (e.g., rational numbers, efficiently stored arrays of integers, etc.), although such additions will often be provided via the standard library instead.</p> <p id="index-5">Some of the type descriptions below contain a paragraph listing ‘special attributes.’ These are attributes that provide access to the implementation and are not intended for general use. Their definition may change in the future.</p> <section id="none"> <h3> +<span class="section-number">3.2.1. </span>None</h3> <p id="index-6">This type has a single value. There is a single object with this value. This object is accessed through the built-in name <code>None</code>. It is used to signify the absence of a value in many situations, e.g., it is returned from functions that don’t explicitly return anything. Its truth value is false.</p> </section> <section id="notimplemented"> <h3> +<span class="section-number">3.2.2. </span>NotImplemented</h3> <p id="index-7">This type has a single value. There is a single object with this value. This object is accessed through the built-in name <code>NotImplemented</code>. Numeric methods and rich comparison methods should return this value if they do not implement the operation for the operands provided. (The interpreter will then try the reflected operation, or some other fallback, depending on the operator.) It should not be evaluated in a boolean context.</p> <p>See <a class="reference internal" href="../library/numbers#implementing-the-arithmetic-operations"><span class="std std-ref">Implementing the arithmetic operations</span></a> for more details.</p> <div class="versionchanged"> <p><span class="versionmodified changed">Changed in version 3.9: </span>Evaluating <code>NotImplemented</code> in a boolean context is deprecated. While it currently evaluates as true, it will emit a <a class="reference internal" href="../library/exceptions#DeprecationWarning" title="DeprecationWarning"><code>DeprecationWarning</code></a>. It will raise a <a class="reference internal" href="../library/exceptions#TypeError" title="TypeError"><code>TypeError</code></a> in a future version of Python.</p> </div> </section> <section id="ellipsis"> <h3> +<span class="section-number">3.2.3. </span>Ellipsis</h3> <p id="index-8">This type has a single value. There is a single object with this value. This object is accessed through the literal <code>...</code> or the built-in name <code>Ellipsis</code>. Its truth value is true.</p> </section> <section id="numbers-number"> <h3> +<span class="section-number">3.2.4. </span><a class="reference internal" href="../library/numbers#numbers.Number" title="numbers.Number"><code>numbers.Number</code></a> +</h3> <p id="index-9">These are created by numeric literals and returned as results by arithmetic operators and arithmetic built-in functions. Numeric objects are immutable; once created their value never changes. Python numbers are of course strongly related to mathematical numbers, but subject to the limitations of numerical representation in computers.</p> <p>The string representations of the numeric classes, computed by <a class="reference internal" href="#object.__repr__" title="object.__repr__"><code>__repr__()</code></a> and <a class="reference internal" href="#object.__str__" title="object.__str__"><code>__str__()</code></a>, have the following properties:</p> <ul class="simple"> <li>They are valid numeric literals which, when passed to their class constructor, produce an object having the value of the original numeric.</li> <li>The representation is in base 10, when possible.</li> <li>Leading zeros, possibly excepting a single zero before a decimal point, are not shown.</li> <li>Trailing zeros, possibly excepting a single zero after a decimal point, are not shown.</li> <li>A sign is shown only when the number is negative.</li> </ul> <p>Python distinguishes between integers, floating point numbers, and complex numbers:</p> <section id="numbers-integral"> <h4> +<span class="section-number">3.2.4.1. </span><a class="reference internal" href="../library/numbers#numbers.Integral" title="numbers.Integral"><code>numbers.Integral</code></a> +</h4> <p id="index-10">These represent elements from the mathematical set of integers (positive and negative).</p> <div class="admonition note"> <p class="admonition-title">Note</p> <p id="index-11">The rules for integer representation are intended to give the most meaningful interpretation of shift and mask operations involving negative integers.</p> </div> <p>There are two types of integers:</p> <dl> <dt> +<code>Integers (int)</code> </dt> +<dd> +<p>These represent numbers in an unlimited range, subject to available (virtual) memory only. For the purpose of shift and mask operations, a binary representation is assumed, and negative numbers are represented in a variant of 2’s complement which gives the illusion of an infinite string of sign bits extending to the left.</p> </dd> <dt> +<code>Booleans (bool)</code> </dt> +<dd> +<p id="index-12">These represent the truth values False and True. The two objects representing the values <code>False</code> and <code>True</code> are the only Boolean objects. The Boolean type is a subtype of the integer type, and Boolean values behave like the values 0 and 1, respectively, in almost all contexts, the exception being that when converted to a string, the strings <code>"False"</code> or <code>"True"</code> are returned, respectively.</p> </dd> </dl> </section> <section id="numbers-real-float"> <h4> +<span class="section-number">3.2.4.2. </span><a class="reference internal" href="../library/numbers#numbers.Real" title="numbers.Real"><code>numbers.Real</code></a> (<a class="reference internal" href="../library/functions#float" title="float"><code>float</code></a>)</h4> <p id="index-13">These represent machine-level double precision floating point numbers. You are at the mercy of the underlying machine architecture (and C or Java implementation) for the accepted range and handling of overflow. Python does not support single-precision floating point numbers; the savings in processor and memory usage that are usually the reason for using these are dwarfed by the overhead of using objects in Python, so there is no reason to complicate the language with two kinds of floating point numbers.</p> </section> <section id="numbers-complex-complex"> <h4> +<span class="section-number">3.2.4.3. </span><a class="reference internal" href="../library/numbers#numbers.Complex" title="numbers.Complex"><code>numbers.Complex</code></a> (<a class="reference internal" href="../library/functions#complex" title="complex"><code>complex</code></a>)</h4> <p id="index-14">These represent complex numbers as a pair of machine-level double precision floating point numbers. The same caveats apply as for floating point numbers. The real and imaginary parts of a complex number <code>z</code> can be retrieved through the read-only attributes <code>z.real</code> and <code>z.imag</code>.</p> </section> </section> <section id="sequences"> <h3> +<span class="section-number">3.2.5. </span>Sequences</h3> <p id="index-15">These represent finite ordered sets indexed by non-negative numbers. The built-in function <a class="reference internal" href="../library/functions#len" title="len"><code>len()</code></a> returns the number of items of a sequence. When the length of a sequence is <em>n</em>, the index set contains the numbers 0, 1, …, <em>n</em>-1. Item <em>i</em> of sequence <em>a</em> is selected by <code>a[i]</code>.</p> <p id="index-16">Sequences also support slicing: <code>a[i:j]</code> selects all items with index <em>k</em> such that <em>i</em> <code><=</code> <em>k</em> <code><</code> <em>j</em>. When used as an expression, a slice is a sequence of the same type. This implies that the index set is renumbered so that it starts at 0.</p> <p>Some sequences also support “extended slicing” with a third “step” parameter: <code>a[i:j:k]</code> selects all items of <em>a</em> with index <em>x</em> where <code>x = i + n*k</code>, <em>n</em> <code>>=</code> <code>0</code> and <em>i</em> <code><=</code> <em>x</em> <code><</code> <em>j</em>.</p> <p>Sequences are distinguished according to their mutability:</p> <section id="immutable-sequences"> <h4> +<span class="section-number">3.2.5.1. </span>Immutable sequences</h4> <p id="index-17">An object of an immutable sequence type cannot change once it is created. (If the object contains references to other objects, these other objects may be mutable and may be changed; however, the collection of objects directly referenced by an immutable object cannot change.)</p> <p>The following types are immutable sequences:</p> <dl id="index-18"> <dt>Strings</dt> +<dd> +<p id="index-19">A string is a sequence of values that represent Unicode code points. All the code points in the range <code>U+0000 - U+10FFFF</code> can be represented in a string. Python doesn’t have a <span class="c-expr sig sig-inline c"><span class="kt">char</span></span> type; instead, every code point in the string is represented as a string object with length <code>1</code>. The built-in function <a class="reference internal" href="../library/functions#ord" title="ord"><code>ord()</code></a> converts a code point from its string form to an integer in the range <code>0 - 10FFFF</code>; <a class="reference internal" href="../library/functions#chr" title="chr"><code>chr()</code></a> converts an integer in the range <code>0 - 10FFFF</code> to the corresponding length <code>1</code> string object. <a class="reference internal" href="../library/stdtypes#str.encode" title="str.encode"><code>str.encode()</code></a> can be used to convert a <a class="reference internal" href="../library/stdtypes#str" title="str"><code>str</code></a> to <a class="reference internal" href="../library/stdtypes#bytes" title="bytes"><code>bytes</code></a> using the given text encoding, and <a class="reference internal" href="../library/stdtypes#bytes.decode" title="bytes.decode"><code>bytes.decode()</code></a> can be used to achieve the opposite.</p> </dd> <dt>Tuples</dt> +<dd> +<p id="index-20">The items of a tuple are arbitrary Python objects. Tuples of two or more items are formed by comma-separated lists of expressions. A tuple of one item (a ‘singleton’) can be formed by affixing a comma to an expression (an expression by itself does not create a tuple, since parentheses must be usable for grouping of expressions). An empty tuple can be formed by an empty pair of parentheses.</p> </dd> <dt>Bytes</dt> +<dd> +<p id="index-21">A bytes object is an immutable array. The items are 8-bit bytes, represented by integers in the range 0 <= x < 256. Bytes literals (like <code>b'abc'</code>) and the built-in <a class="reference internal" href="../library/stdtypes#bytes" title="bytes"><code>bytes()</code></a> constructor can be used to create bytes objects. Also, bytes objects can be decoded to strings via the <a class="reference internal" href="../library/stdtypes#bytes.decode" title="bytes.decode"><code>decode()</code></a> method.</p> </dd> </dl> </section> <section id="mutable-sequences"> <h4> +<span class="section-number">3.2.5.2. </span>Mutable sequences</h4> <p id="index-22">Mutable sequences can be changed after they are created. The subscription and slicing notations can be used as the target of assignment and <a class="reference internal" href="simple_stmts#del"><code>del</code></a> (delete) statements.</p> <div class="admonition note"> <p class="admonition-title">Note</p> <span class="target" id="index-23"></span><p id="index-24">The <a class="reference internal" href="../library/collections#module-collections" title="collections: Container datatypes"><code>collections</code></a> and <a class="reference internal" href="../library/array#module-array" title="array: Space efficient arrays of uniformly typed numeric values."><code>array</code></a> module provide additional examples of mutable sequence types.</p> </div> <p>There are currently two intrinsic mutable sequence types:</p> <dl> <dt>Lists</dt> +<dd> +<p id="index-25">The items of a list are arbitrary Python objects. Lists are formed by placing a comma-separated list of expressions in square brackets. (Note that there are no special cases needed to form lists of length 0 or 1.)</p> </dd> <dt>Byte Arrays</dt> +<dd> +<p id="index-26">A bytearray object is a mutable array. They are created by the built-in <a class="reference internal" href="../library/stdtypes#bytearray" title="bytearray"><code>bytearray()</code></a> constructor. Aside from being mutable (and hence unhashable), byte arrays otherwise provide the same interface and functionality as immutable <a class="reference internal" href="../library/stdtypes#bytes" title="bytes"><code>bytes</code></a> objects.</p> </dd> </dl> </section> </section> <section id="set-types"> <h3> +<span class="section-number">3.2.6. </span>Set types</h3> <p id="index-27">These represent unordered, finite sets of unique, immutable objects. As such, they cannot be indexed by any subscript. However, they can be iterated over, and the built-in function <a class="reference internal" href="../library/functions#len" title="len"><code>len()</code></a> returns the number of items in a set. Common uses for sets are fast membership testing, removing duplicates from a sequence, and computing mathematical operations such as intersection, union, difference, and symmetric difference.</p> <p>For set elements, the same immutability rules apply as for dictionary keys. Note that numeric types obey the normal rules for numeric comparison: if two numbers compare equal (e.g., <code>1</code> and <code>1.0</code>), only one of them can be contained in a set.</p> <p>There are currently two intrinsic set types:</p> <dl> <dt>Sets</dt> +<dd> +<p id="index-28">These represent a mutable set. They are created by the built-in <a class="reference internal" href="../library/stdtypes#set" title="set"><code>set()</code></a> constructor and can be modified afterwards by several methods, such as <code>add()</code>.</p> </dd> <dt>Frozen sets</dt> +<dd> +<p id="index-29">These represent an immutable set. They are created by the built-in <a class="reference internal" href="../library/stdtypes#frozenset" title="frozenset"><code>frozenset()</code></a> constructor. As a frozenset is immutable and <a class="reference internal" href="../glossary#term-hashable"><span class="xref std std-term">hashable</span></a>, it can be used again as an element of another set, or as a dictionary key.</p> </dd> </dl> </section> <section id="mappings"> <h3> +<span class="section-number">3.2.7. </span>Mappings</h3> <p id="index-30">These represent finite sets of objects indexed by arbitrary index sets. The subscript notation <code>a[k]</code> selects the item indexed by <code>k</code> from the mapping <code>a</code>; this can be used in expressions and as the target of assignments or <a class="reference internal" href="simple_stmts#del"><code>del</code></a> statements. The built-in function <a class="reference internal" href="../library/functions#len" title="len"><code>len()</code></a> returns the number of items in a mapping.</p> <p>There is currently a single intrinsic mapping type:</p> <section id="dictionaries"> <h4> +<span class="section-number">3.2.7.1. </span>Dictionaries</h4> <p id="index-31">These represent finite sets of objects indexed by nearly arbitrary values. The only types of values not acceptable as keys are values containing lists or dictionaries or other mutable types that are compared by value rather than by object identity, the reason being that the efficient implementation of dictionaries requires a key’s hash value to remain constant. Numeric types used for keys obey the normal rules for numeric comparison: if two numbers compare equal (e.g., <code>1</code> and <code>1.0</code>) then they can be used interchangeably to index the same dictionary entry.</p> <p>Dictionaries preserve insertion order, meaning that keys will be produced in the same order they were added sequentially over the dictionary. Replacing an existing key does not change the order, however removing a key and re-inserting it will add it to the end instead of keeping its old place.</p> <p>Dictionaries are mutable; they can be created by the <code>{...}</code> notation (see section <a class="reference internal" href="expressions#dict"><span class="std std-ref">Dictionary displays</span></a>).</p> <p id="index-32">The extension modules <a class="reference internal" href="../library/dbm#module-dbm.ndbm" title='dbm.ndbm: The standard "database" interface, based on ndbm. (Unix)'><code>dbm.ndbm</code></a> and <a class="reference internal" href="../library/dbm#module-dbm.gnu" title="dbm.gnu: GNU's reinterpretation of dbm. (Unix)"><code>dbm.gnu</code></a> provide additional examples of mapping types, as does the <a class="reference internal" href="../library/collections#module-collections" title="collections: Container datatypes"><code>collections</code></a> module.</p> <div class="versionchanged"> <p><span class="versionmodified changed">Changed in version 3.7: </span>Dictionaries did not preserve insertion order in versions of Python before 3.6. In CPython 3.6, insertion order was preserved, but it was considered an implementation detail at that time rather than a language guarantee.</p> </div> </section> </section> <section id="callable-types"> <h3> +<span class="section-number">3.2.8. </span>Callable types</h3> <p id="index-33">These are the types to which the function call operation (see section <a class="reference internal" href="expressions#calls"><span class="std std-ref">Calls</span></a>) can be applied:</p> <section id="user-defined-functions"> <span id="user-defined-funcs"></span><h4> +<span class="section-number">3.2.8.1. </span>User-defined functions</h4> <p id="index-34">A user-defined function object is created by a function definition (see section <a class="reference internal" href="compound_stmts#function"><span class="std std-ref">Function definitions</span></a>). It should be called with an argument list containing the same number of items as the function’s formal parameter list.</p> <section id="special-read-only-attributes"> <h5> +<span class="section-number">3.2.8.1.1. </span>Special read-only attributes</h5> <table class="docutils align-default" id="index-35"> <thead> <tr> +<th class="head"><p>Attribute</p></th> <th class="head"><p>Meaning</p></th> </tr> </thead> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="function.__globals__"> +<code>function.__globals__</code> </dt> <dd></dd> +</dl> </td> <td><p>A reference to the <a class="reference internal" href="../library/stdtypes#dict" title="dict"><code>dictionary</code></a> that holds the function’s <a class="reference internal" href="executionmodel#naming"><span class="std std-ref">global variables</span></a> – the global namespace of the module in which the function was defined.</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="function.__closure__"> +<code>function.__closure__</code> </dt> <dd></dd> +</dl> </td> <td> +<p><code>None</code> or a <a class="reference internal" href="../library/stdtypes#tuple" title="tuple"><code>tuple</code></a> of cells that contain bindings for the function’s free variables.</p> <p>A cell object has the attribute <code>cell_contents</code>. This can be used to get the value of the cell, as well as set the value.</p> </td> </tr> </table> </section> <section id="special-writable-attributes"> <h5> +<span class="section-number">3.2.8.1.2. </span>Special writable attributes</h5> <p id="index-36">Most of these attributes check the type of the assigned value:</p> <table class="docutils align-default"> <thead> <tr> +<th class="head"><p>Attribute</p></th> <th class="head"><p>Meaning</p></th> </tr> </thead> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="function.__doc__"> +<code>function.__doc__</code> </dt> <dd></dd> +</dl> </td> <td><p>The function’s documentation string, or <code>None</code> if unavailable. Not inherited by subclasses.</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="function.__name__"> +<code>function.__name__</code> </dt> <dd></dd> +</dl> </td> <td><p>The function’s name. See also: <a class="reference internal" href="../library/stdtypes#definition.__name__" title="definition.__name__"><code>__name__ attributes</code></a>.</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="function.__qualname__"> +<code>function.__qualname__</code> </dt> <dd></dd> +</dl> </td> <td> +<p>The function’s <a class="reference internal" href="../glossary#term-qualified-name"><span class="xref std std-term">qualified name</span></a>. See also: <a class="reference internal" href="../library/stdtypes#definition.__qualname__" title="definition.__qualname__"><code>__qualname__ attributes</code></a>.</p> <div class="versionadded"> <p><span class="versionmodified added">New in version 3.3.</span></p> </div> </td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="function.__module__"> +<code>function.__module__</code> </dt> <dd></dd> +</dl> </td> <td><p>The name of the module the function was defined in, or <code>None</code> if unavailable.</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="function.__defaults__"> +<code>function.__defaults__</code> </dt> <dd></dd> +</dl> </td> <td><p>A <a class="reference internal" href="../library/stdtypes#tuple" title="tuple"><code>tuple</code></a> containing default <a class="reference internal" href="../glossary#term-parameter"><span class="xref std std-term">parameter</span></a> values for those parameters that have defaults, or <code>None</code> if no parameters have a default value.</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="function.__code__"> +<code>function.__code__</code> </dt> <dd></dd> +</dl> </td> <td><p>The <a class="reference internal" href="#code-objects"><span class="std std-ref">code object</span></a> representing the compiled function body.</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="function.__dict__"> +<code>function.__dict__</code> </dt> <dd></dd> +</dl> </td> <td><p>The namespace supporting arbitrary function attributes. See also: <a class="reference internal" href="../library/stdtypes#object.__dict__" title="object.__dict__"><code>__dict__ attributes</code></a>.</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="function.__annotations__"> +<code>function.__annotations__</code> </dt> <dd></dd> +</dl> </td> <td><p>A <a class="reference internal" href="../library/stdtypes#dict" title="dict"><code>dictionary</code></a> containing annotations of <a class="reference internal" href="../glossary#term-parameter"><span class="xref std std-term">parameters</span></a>. The keys of the dictionary are the parameter names, and <code>'return'</code> for the return annotation, if provided. See also: <a class="reference internal" href="../howto/annotations#annotations-howto"><span class="std std-ref">Annotations Best Practices</span></a>.</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="function.__kwdefaults__"> +<code>function.__kwdefaults__</code> </dt> <dd></dd> +</dl> </td> <td><p>A <a class="reference internal" href="../library/stdtypes#dict" title="dict"><code>dictionary</code></a> containing defaults for keyword-only <a class="reference internal" href="../glossary#term-parameter"><span class="xref std std-term">parameters</span></a>.</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="function.__type_params__"> +<code>function.__type_params__</code> </dt> <dd></dd> +</dl> </td> <td> +<p>A <a class="reference internal" href="../library/stdtypes#tuple" title="tuple"><code>tuple</code></a> containing the <a class="reference internal" href="compound_stmts#type-params"><span class="std std-ref">type parameters</span></a> of a <a class="reference internal" href="compound_stmts#generic-functions"><span class="std std-ref">generic function</span></a>.</p> <div class="versionadded"> <p><span class="versionmodified added">New in version 3.12.</span></p> </div> </td> </tr> </table> <p>Function objects also support getting and setting arbitrary attributes, which can be used, for example, to attach metadata to functions. Regular attribute dot-notation is used to get and set such attributes.</p> <div class="impl-detail compound"> <p><strong>CPython implementation detail:</strong> CPython’s current implementation only supports function attributes on user-defined functions. Function attributes on <a class="reference internal" href="#builtin-functions"><span class="std std-ref">built-in functions</span></a> may be supported in the future.</p> </div> <p>Additional information about a function’s definition can be retrieved from its <a class="reference internal" href="#code-objects"><span class="std std-ref">code object</span></a> (accessible via the <a class="reference internal" href="#function.__code__" title="function.__code__"><code>__code__</code></a> attribute).</p> </section> </section> <section id="instance-methods"> <span id="id2"></span><h4> +<span class="section-number">3.2.8.2. </span>Instance methods</h4> <p id="index-37">An instance method object combines a class, a class instance and any callable object (normally a user-defined function).</p> <p id="index-38">Special read-only attributes:</p> <table class="docutils align-default"> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="method.__self__"> +<code>method.__self__</code> </dt> <dd></dd> +</dl> </td> <td><p>Refers to the class instance object to which the method is <a class="reference internal" href="#method-binding"><span class="std std-ref">bound</span></a></p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="method.__func__"> +<code>method.__func__</code> </dt> <dd></dd> +</dl> </td> <td><p>Refers to the original <a class="reference internal" href="#user-defined-funcs"><span class="std std-ref">function object</span></a></p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="method.__doc__"> +<code>method.__doc__</code> </dt> <dd></dd> +</dl> </td> <td><p>The method’s documentation (same as <a class="reference internal" href="#function.__doc__" title="function.__doc__"><code>method.__func__.__doc__</code></a>). A <a class="reference internal" href="../library/stdtypes#str" title="str"><code>string</code></a> if the original function had a docstring, else <code>None</code>.</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="method.__name__"> +<code>method.__name__</code> </dt> <dd></dd> +</dl> </td> <td><p>The name of the method (same as <a class="reference internal" href="#function.__name__" title="function.__name__"><code>method.__func__.__name__</code></a>)</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="method.__module__"> +<code>method.__module__</code> </dt> <dd></dd> +</dl> </td> <td><p>The name of the module the method was defined in, or <code>None</code> if unavailable.</p></td> </tr> </table> <p>Methods also support accessing (but not setting) the arbitrary function attributes on the underlying <a class="reference internal" href="#user-defined-funcs"><span class="std std-ref">function object</span></a>.</p> <p>User-defined method objects may be created when getting an attribute of a class (perhaps via an instance of that class), if that attribute is a user-defined <a class="reference internal" href="#user-defined-funcs"><span class="std std-ref">function object</span></a> or a <a class="reference internal" href="../library/functions#classmethod" title="classmethod"><code>classmethod</code></a> object.</p> <p id="method-binding">When an instance method object is created by retrieving a user-defined <a class="reference internal" href="#user-defined-funcs"><span class="std std-ref">function object</span></a> from a class via one of its instances, its <a class="reference internal" href="#method.__self__" title="method.__self__"><code>__self__</code></a> attribute is the instance, and the method object is said to be <em>bound</em>. The new method’s <a class="reference internal" href="#method.__func__" title="method.__func__"><code>__func__</code></a> attribute is the original function object.</p> <p>When an instance method object is created by retrieving a <a class="reference internal" href="../library/functions#classmethod" title="classmethod"><code>classmethod</code></a> object from a class or instance, its <a class="reference internal" href="#method.__self__" title="method.__self__"><code>__self__</code></a> attribute is the class itself, and its <a class="reference internal" href="#method.__func__" title="method.__func__"><code>__func__</code></a> attribute is the function object underlying the class method.</p> <p>When an instance method object is called, the underlying function (<a class="reference internal" href="#method.__func__" title="method.__func__"><code>__func__</code></a>) is called, inserting the class instance (<a class="reference internal" href="#method.__self__" title="method.__self__"><code>__self__</code></a>) in front of the argument list. For instance, when <code>C</code> is a class which contains a definition for a function <code>f()</code>, and <code>x</code> is an instance of <code>C</code>, calling <code>x.f(1)</code> is equivalent to calling <code>C.f(x, 1)</code>.</p> <p>When an instance method object is derived from a <a class="reference internal" href="../library/functions#classmethod" title="classmethod"><code>classmethod</code></a> object, the “class instance” stored in <a class="reference internal" href="#method.__self__" title="method.__self__"><code>__self__</code></a> will actually be the class itself, so that calling either <code>x.f(1)</code> or <code>C.f(1)</code> is equivalent to calling <code>f(C,1)</code> where <code>f</code> is the underlying function.</p> <p>Note that the transformation from <a class="reference internal" href="#user-defined-funcs"><span class="std std-ref">function object</span></a> to instance method object happens each time the attribute is retrieved from the instance. In some cases, a fruitful optimization is to assign the attribute to a local variable and call that local variable. Also notice that this transformation only happens for user-defined functions; other callable objects (and all non-callable objects) are retrieved without transformation. It is also important to note that user-defined functions which are attributes of a class instance are not converted to bound methods; this <em>only</em> happens when the function is an attribute of the class.</p> </section> <section id="generator-functions"> <h4> +<span class="section-number">3.2.8.3. </span>Generator functions</h4> <p id="index-39">A function or method which uses the <a class="reference internal" href="simple_stmts#yield"><code>yield</code></a> statement (see section <a class="reference internal" href="simple_stmts#yield"><span class="std std-ref">The yield statement</span></a>) is called a <em class="dfn">generator function</em>. Such a function, when called, always returns an <a class="reference internal" href="../glossary#term-iterator"><span class="xref std std-term">iterator</span></a> object which can be used to execute the body of the function: calling the iterator’s <a class="reference internal" href="../library/stdtypes#iterator.__next__" title="iterator.__next__"><code>iterator.__next__()</code></a> method will cause the function to execute until it provides a value using the <code>yield</code> statement. When the function executes a <a class="reference internal" href="simple_stmts#return"><code>return</code></a> statement or falls off the end, a <a class="reference internal" href="../library/exceptions#StopIteration" title="StopIteration"><code>StopIteration</code></a> exception is raised and the iterator will have reached the end of the set of values to be returned.</p> </section> <section id="coroutine-functions"> <h4> +<span class="section-number">3.2.8.4. </span>Coroutine functions</h4> <p id="index-40">A function or method which is defined using <a class="reference internal" href="compound_stmts#async-def"><code>async def</code></a> is called a <em class="dfn">coroutine function</em>. Such a function, when called, returns a <a class="reference internal" href="../glossary#term-coroutine"><span class="xref std std-term">coroutine</span></a> object. It may contain <a class="reference internal" href="expressions#await"><code>await</code></a> expressions, as well as <a class="reference internal" href="compound_stmts#async-with"><code>async with</code></a> and <a class="reference internal" href="compound_stmts#async-for"><code>async for</code></a> statements. See also the <a class="reference internal" href="#coroutine-objects"><span class="std std-ref">Coroutine Objects</span></a> section.</p> </section> <section id="asynchronous-generator-functions"> <h4> +<span class="section-number">3.2.8.5. </span>Asynchronous generator functions</h4> <p id="index-41">A function or method which is defined using <a class="reference internal" href="compound_stmts#async-def"><code>async def</code></a> and which uses the <a class="reference internal" href="simple_stmts#yield"><code>yield</code></a> statement is called a <em class="dfn">asynchronous generator function</em>. Such a function, when called, returns an <a class="reference internal" href="../glossary#term-asynchronous-iterator"><span class="xref std std-term">asynchronous iterator</span></a> object which can be used in an <a class="reference internal" href="compound_stmts#async-for"><code>async for</code></a> statement to execute the body of the function.</p> <p>Calling the asynchronous iterator’s <a class="reference internal" href="#object.__anext__" title="object.__anext__"><code>aiterator.__anext__</code></a> method will return an <a class="reference internal" href="../glossary#term-awaitable"><span class="xref std std-term">awaitable</span></a> which when awaited will execute until it provides a value using the <a class="reference internal" href="simple_stmts#yield"><code>yield</code></a> expression. When the function executes an empty <a class="reference internal" href="simple_stmts#return"><code>return</code></a> statement or falls off the end, a <a class="reference internal" href="../library/exceptions#StopAsyncIteration" title="StopAsyncIteration"><code>StopAsyncIteration</code></a> exception is raised and the asynchronous iterator will have reached the end of the set of values to be yielded.</p> </section> <section id="built-in-functions"> <span id="builtin-functions"></span><h4> +<span class="section-number">3.2.8.6. </span>Built-in functions</h4> <p id="index-42">A built-in function object is a wrapper around a C function. Examples of built-in functions are <a class="reference internal" href="../library/functions#len" title="len"><code>len()</code></a> and <a class="reference internal" href="../library/math#math.sin" title="math.sin"><code>math.sin()</code></a> (<a class="reference internal" href="../library/math#module-math" title="math: Mathematical functions (sin() etc.)."><code>math</code></a> is a standard built-in module). The number and type of the arguments are determined by the C function. Special read-only attributes:</p> <ul class="simple"> <li> +<code>__doc__</code> is the function’s documentation string, or <code>None</code> if unavailable. See <a class="reference internal" href="#function.__doc__" title="function.__doc__"><code>function.__doc__</code></a>.</li> <li> +<code>__name__</code> is the function’s name. See <a class="reference internal" href="#function.__name__" title="function.__name__"><code>function.__name__</code></a>.</li> <li> +<code>__self__</code> is set to <code>None</code> (but see the next item).</li> <li> +<code>__module__</code> is the name of the module the function was defined in or <code>None</code> if unavailable. See <a class="reference internal" href="#function.__module__" title="function.__module__"><code>function.__module__</code></a>.</li> </ul> </section> <section id="built-in-methods"> <span id="builtin-methods"></span><h4> +<span class="section-number">3.2.8.7. </span>Built-in methods</h4> <p id="index-43">This is really a different disguise of a built-in function, this time containing an object passed to the C function as an implicit extra argument. An example of a built-in method is <code>alist.append()</code>, assuming <em>alist</em> is a list object. In this case, the special read-only attribute <code>__self__</code> is set to the object denoted by <em>alist</em>. (The attribute has the same semantics as it does with <a class="reference internal" href="#method.__self__" title="method.__self__"><code>other instance methods</code></a>.)</p> </section> <section id="classes"> <h4> +<span class="section-number">3.2.8.8. </span>Classes</h4> <p>Classes are callable. These objects normally act as factories for new instances of themselves, but variations are possible for class types that override <a class="reference internal" href="#object.__new__" title="object.__new__"><code>__new__()</code></a>. The arguments of the call are passed to <code>__new__()</code> and, in the typical case, to <a class="reference internal" href="#object.__init__" title="object.__init__"><code>__init__()</code></a> to initialize the new instance.</p> </section> <section id="class-instances"> <h4> +<span class="section-number">3.2.8.9. </span>Class Instances</h4> <p>Instances of arbitrary classes can be made callable by defining a <a class="reference internal" href="#object.__call__" title="object.__call__"><code>__call__()</code></a> method in their class.</p> </section> </section> <section id="modules"> <h3> +<span class="section-number">3.2.9. </span>Modules</h3> <p id="index-44">Modules are a basic organizational unit of Python code, and are created by the <a class="reference internal" href="import#importsystem"><span class="std std-ref">import system</span></a> as invoked either by the <a class="reference internal" href="simple_stmts#import"><code>import</code></a> statement, or by calling functions such as <a class="reference internal" href="../library/importlib#importlib.import_module" title="importlib.import_module"><code>importlib.import_module()</code></a> and built-in <a class="reference internal" href="../library/functions#import__" title="__import__"><code>__import__()</code></a>. A module object has a namespace implemented by a <a class="reference internal" href="../library/stdtypes#dict" title="dict"><code>dictionary</code></a> object (this is the dictionary referenced by the <a class="reference internal" href="#function.__globals__" title="function.__globals__"><code>__globals__</code></a> attribute of functions defined in the module). Attribute references are translated to lookups in this dictionary, e.g., <code>m.x</code> is equivalent to <code>m.__dict__["x"]</code>. A module object does not contain the code object used to initialize the module (since it isn’t needed once the initialization is done).</p> <p>Attribute assignment updates the module’s namespace dictionary, e.g., <code>m.x = 1</code> is equivalent to <code>m.__dict__["x"] = 1</code>.</p> <p id="index-45">Predefined (writable) attributes:</p> <dl class="simple"> <dt> +<code></code> <a class="reference internal" href="import#name__" title="__name__"><code>__name__</code></a> +</dt> +<dd> +<p>The module’s name.</p> </dd> <dt> +<code>__doc__</code> </dt> +<dd> +<p>The module’s documentation string, or <code>None</code> if unavailable.</p> </dd> <dt> +<code></code> <a class="reference internal" href="import#file__" title="__file__"><code>__file__</code></a> +</dt> +<dd> +<p>The pathname of the file from which the module was loaded, if it was loaded from a file. The <a class="reference internal" href="import#file__" title="__file__"><code>__file__</code></a> attribute may be missing for certain types of modules, such as C modules that are statically linked into the interpreter. For extension modules loaded dynamically from a shared library, it’s the pathname of the shared library file.</p> </dd> <dt> +<code>__annotations__</code> </dt> +<dd> +<p>A dictionary containing <a class="reference internal" href="../glossary#term-variable-annotation"><span class="xref std std-term">variable annotations</span></a> collected during module body execution. For best practices on working with <code>__annotations__</code>, please see <a class="reference internal" href="../howto/annotations#annotations-howto"><span class="std std-ref">Annotations Best Practices</span></a>.</p> </dd> </dl> <p id="index-46">Special read-only attribute: <a class="reference internal" href="../library/stdtypes#object.__dict__" title="object.__dict__"><code>__dict__</code></a> is the module’s namespace as a dictionary object.</p> <div class="impl-detail compound"> <p><strong>CPython implementation detail:</strong> Because of the way CPython clears module dictionaries, the module dictionary will be cleared when the module falls out of scope even if the dictionary still has live references. To avoid this, copy the dictionary or keep the module around while using its dictionary directly.</p> </div> </section> <section id="custom-classes"> <h3> +<span class="section-number">3.2.10. </span>Custom classes</h3> <p>Custom class types are typically created by class definitions (see section <a class="reference internal" href="compound_stmts#class"><span class="std std-ref">Class definitions</span></a>). A class has a namespace implemented by a dictionary object. Class attribute references are translated to lookups in this dictionary, e.g., <code>C.x</code> is translated to <code>C.__dict__["x"]</code> (although there are a number of hooks which allow for other means of locating attributes). When the attribute name is not found there, the attribute search continues in the base classes. This search of the base classes uses the C3 method resolution order which behaves correctly even in the presence of ‘diamond’ inheritance structures where there are multiple inheritance paths leading back to a common ancestor. Additional details on the C3 MRO used by Python can be found in the documentation accompanying the 2.3 release at <a class="reference external" href="https://www.python.org/download/releases/2.3/mro/">https://www.python.org/download/releases/2.3/mro/</a>.</p> <p id="index-47">When a class attribute reference (for class <code>C</code>, say) would yield a class method object, it is transformed into an instance method object whose <a class="reference internal" href="#method.__self__" title="method.__self__"><code>__self__</code></a> attribute is <code>C</code>. When it would yield a <a class="reference internal" href="../library/functions#staticmethod" title="staticmethod"><code>staticmethod</code></a> object, it is transformed into the object wrapped by the static method object. See section <a class="reference internal" href="#descriptors"><span class="std std-ref">Implementing Descriptors</span></a> for another way in which attributes retrieved from a class may differ from those actually contained in its <a class="reference internal" href="../library/stdtypes#object.__dict__" title="object.__dict__"><code>__dict__</code></a>.</p> <p id="index-48">Class attribute assignments update the class’s dictionary, never the dictionary of a base class.</p> <p id="index-49">A class object can be called (see above) to yield a class instance (see below).</p> <p id="index-50">Special attributes:</p> <dl class="simple"> <dt> +<code></code> <a class="reference internal" href="../library/stdtypes#definition.__name__" title="definition.__name__"><code>__name__</code></a> +</dt> +<dd> +<p>The class name.</p> </dd> <dt> +<code>__module__</code> </dt> +<dd> +<p>The name of the module in which the class was defined.</p> </dd> <dt> +<code></code> <a class="reference internal" href="../library/stdtypes#object.__dict__" title="object.__dict__"><code>__dict__</code></a> +</dt> +<dd> +<p>The dictionary containing the class’s namespace.</p> </dd> <dt> +<code></code> <a class="reference internal" href="../library/stdtypes#class.__bases__" title="class.__bases__"><code>__bases__</code></a> +</dt> +<dd> +<p>A tuple containing the base classes, in the order of their occurrence in the base class list.</p> </dd> <dt> +<code>__doc__</code> </dt> +<dd> +<p>The class’s documentation string, or <code>None</code> if undefined.</p> </dd> <dt> +<code>__annotations__</code> </dt> +<dd> +<p>A dictionary containing <a class="reference internal" href="../glossary#term-variable-annotation"><span class="xref std std-term">variable annotations</span></a> collected during class body execution. For best practices on working with <code>__annotations__</code>, please see <a class="reference internal" href="../howto/annotations#annotations-howto"><span class="std std-ref">Annotations Best Practices</span></a>.</p> </dd> <dt> +<code>__type_params__</code> </dt> +<dd> +<p>A tuple containing the <a class="reference internal" href="compound_stmts#type-params"><span class="std std-ref">type parameters</span></a> of a <a class="reference internal" href="compound_stmts#generic-classes"><span class="std std-ref">generic class</span></a>.</p> </dd> </dl> </section> <section id="id3"> <h3> +<span class="section-number">3.2.11. </span>Class instances</h3> <p id="index-51">A class instance is created by calling a class object (see above). A class instance has a namespace implemented as a dictionary which is the first place in which attribute references are searched. When an attribute is not found there, and the instance’s class has an attribute by that name, the search continues with the class attributes. If a class attribute is found that is a user-defined function object, it is transformed into an instance method object whose <a class="reference internal" href="#method.__self__" title="method.__self__"><code>__self__</code></a> attribute is the instance. Static method and class method objects are also transformed; see above under “Classes”. See section <a class="reference internal" href="#descriptors"><span class="std std-ref">Implementing Descriptors</span></a> for another way in which attributes of a class retrieved via its instances may differ from the objects actually stored in the class’s <a class="reference internal" href="../library/stdtypes#object.__dict__" title="object.__dict__"><code>__dict__</code></a>. If no class attribute is found, and the object’s class has a <a class="reference internal" href="#object.__getattr__" title="object.__getattr__"><code>__getattr__()</code></a> method, that is called to satisfy the lookup.</p> <p id="index-52">Attribute assignments and deletions update the instance’s dictionary, never a class’s dictionary. If the class has a <a class="reference internal" href="#object.__setattr__" title="object.__setattr__"><code>__setattr__()</code></a> or <a class="reference internal" href="#object.__delattr__" title="object.__delattr__"><code>__delattr__()</code></a> method, this is called instead of updating the instance dictionary directly.</p> <p id="index-53">Class instances can pretend to be numbers, sequences, or mappings if they have methods with certain special names. See section <a class="reference internal" href="#specialnames"><span class="std std-ref">Special method names</span></a>.</p> <p id="index-54">Special attributes: <a class="reference internal" href="../library/stdtypes#object.__dict__" title="object.__dict__"><code>__dict__</code></a> is the attribute dictionary; <a class="reference internal" href="../library/stdtypes#instance.__class__" title="instance.__class__"><code>__class__</code></a> is the instance’s class.</p> </section> <section id="i-o-objects-also-known-as-file-objects"> <h3> +<span class="section-number">3.2.12. </span>I/O objects (also known as file objects)</h3> <p id="index-55">A <a class="reference internal" href="../glossary#term-file-object"><span class="xref std std-term">file object</span></a> represents an open file. Various shortcuts are available to create file objects: the <a class="reference internal" href="../library/functions#open" title="open"><code>open()</code></a> built-in function, and also <a class="reference internal" href="../library/os#os.popen" title="os.popen"><code>os.popen()</code></a>, <a class="reference internal" href="../library/os#os.fdopen" title="os.fdopen"><code>os.fdopen()</code></a>, and the <a class="reference internal" href="../library/socket#socket.socket.makefile" title="socket.socket.makefile"><code>makefile()</code></a> method of socket objects (and perhaps by other functions or methods provided by extension modules).</p> <p>The objects <code>sys.stdin</code>, <code>sys.stdout</code> and <code>sys.stderr</code> are initialized to file objects corresponding to the interpreter’s standard input, output and error streams; they are all open in text mode and therefore follow the interface defined by the <a class="reference internal" href="../library/io#io.TextIOBase" title="io.TextIOBase"><code>io.TextIOBase</code></a> abstract class.</p> </section> <section id="internal-types"> <h3> +<span class="section-number">3.2.13. </span>Internal types</h3> <p id="index-56">A few types used internally by the interpreter are exposed to the user. Their definitions may change with future versions of the interpreter, but they are mentioned here for completeness.</p> <section id="code-objects"> <span id="id4"></span><h4> +<span class="section-number">3.2.13.1. </span>Code objects</h4> <p id="index-57">Code objects represent <em>byte-compiled</em> executable Python code, or <a class="reference internal" href="../glossary#term-bytecode"><span class="xref std std-term">bytecode</span></a>. The difference between a code object and a function object is that the function object contains an explicit reference to the function’s globals (the module in which it was defined), while a code object contains no context; also the default argument values are stored in the function object, not in the code object (because they represent values calculated at run-time). Unlike function objects, code objects are immutable and contain no references (directly or indirectly) to mutable objects.</p> <section id="index-58"> <span id="id5"></span><h5> +<span class="section-number">3.2.13.1.1. </span>Special read-only attributes</h5> <table class="docutils align-default"> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="codeobject.co_name"> +<code>codeobject.co_name</code> </dt> <dd></dd> +</dl> </td> <td><p>The function name</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="codeobject.co_qualname"> +<code>codeobject.co_qualname</code> </dt> <dd></dd> +</dl> </td> <td><p>The fully qualified function name</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="codeobject.co_argcount"> +<code>codeobject.co_argcount</code> </dt> <dd></dd> +</dl> </td> <td><p>The total number of positional <a class="reference internal" href="../glossary#term-parameter"><span class="xref std std-term">parameters</span></a> (including positional-only parameters and parameters with default values) that the function has</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="codeobject.co_posonlyargcount"> +<code>codeobject.co_posonlyargcount</code> </dt> <dd></dd> +</dl> </td> <td><p>The number of positional-only <a class="reference internal" href="../glossary#term-parameter"><span class="xref std std-term">parameters</span></a> (including arguments with default values) that the function has</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="codeobject.co_kwonlyargcount"> +<code>codeobject.co_kwonlyargcount</code> </dt> <dd></dd> +</dl> </td> <td><p>The number of keyword-only <a class="reference internal" href="../glossary#term-parameter"><span class="xref std std-term">parameters</span></a> (including arguments with default values) that the function has</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="codeobject.co_nlocals"> +<code>codeobject.co_nlocals</code> </dt> <dd></dd> +</dl> </td> <td><p>The number of <a class="reference internal" href="executionmodel#naming"><span class="std std-ref">local variables</span></a> used by the function (including parameters)</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="codeobject.co_varnames"> +<code>codeobject.co_varnames</code> </dt> <dd></dd> +</dl> </td> <td><p>A <a class="reference internal" href="../library/stdtypes#tuple" title="tuple"><code>tuple</code></a> containing the names of the local variables in the function (starting with the parameter names)</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="codeobject.co_cellvars"> +<code>codeobject.co_cellvars</code> </dt> <dd></dd> +</dl> </td> <td><p>A <a class="reference internal" href="../library/stdtypes#tuple" title="tuple"><code>tuple</code></a> containing the names of <a class="reference internal" href="executionmodel#naming"><span class="std std-ref">local variables</span></a> that are referenced by nested functions inside the function</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="codeobject.co_freevars"> +<code>codeobject.co_freevars</code> </dt> <dd></dd> +</dl> </td> <td><p>A <a class="reference internal" href="../library/stdtypes#tuple" title="tuple"><code>tuple</code></a> containing the names of free variables in the function</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="codeobject.co_code"> +<code>codeobject.co_code</code> </dt> <dd></dd> +</dl> </td> <td><p>A string representing the sequence of <a class="reference internal" href="../glossary#term-bytecode"><span class="xref std std-term">bytecode</span></a> instructions in the function</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="codeobject.co_consts"> +<code>codeobject.co_consts</code> </dt> <dd></dd> +</dl> </td> <td><p>A <a class="reference internal" href="../library/stdtypes#tuple" title="tuple"><code>tuple</code></a> containing the literals used by the <a class="reference internal" href="../glossary#term-bytecode"><span class="xref std std-term">bytecode</span></a> in the function</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="codeobject.co_names"> +<code>codeobject.co_names</code> </dt> <dd></dd> +</dl> </td> <td><p>A <a class="reference internal" href="../library/stdtypes#tuple" title="tuple"><code>tuple</code></a> containing the names used by the <a class="reference internal" href="../glossary#term-bytecode"><span class="xref std std-term">bytecode</span></a> in the function</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="codeobject.co_filename"> +<code>codeobject.co_filename</code> </dt> <dd></dd> +</dl> </td> <td><p>The name of the file from which the code was compiled</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="codeobject.co_firstlineno"> +<code>codeobject.co_firstlineno</code> </dt> <dd></dd> +</dl> </td> <td><p>The line number of the first line of the function</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="codeobject.co_lnotab"> +<code>codeobject.co_lnotab</code> </dt> <dd></dd> +</dl> </td> <td> +<p>A string encoding the mapping from <a class="reference internal" href="../glossary#term-bytecode"><span class="xref std std-term">bytecode</span></a> offsets to line numbers. For details, see the source code of the interpreter.</p> <div class="deprecated"> <p><span class="versionmodified deprecated">Deprecated since version 3.12: </span>This attribute of code objects is deprecated, and may be removed in Python 3.14.</p> </div> </td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="codeobject.co_stacksize"> +<code>codeobject.co_stacksize</code> </dt> <dd></dd> +</dl> </td> <td><p>The required stack size of the code object</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="codeobject.co_flags"> +<code>codeobject.co_flags</code> </dt> <dd></dd> +</dl> </td> <td><p>An <a class="reference internal" href="../library/functions#int" title="int"><code>integer</code></a> encoding a number of flags for the interpreter.</p></td> </tr> </table> <p id="index-59">The following flag bits are defined for <a class="reference internal" href="#codeobject.co_flags" title="codeobject.co_flags"><code>co_flags</code></a>: bit <code>0x04</code> is set if the function uses the <code>*arguments</code> syntax to accept an arbitrary number of positional arguments; bit <code>0x08</code> is set if the function uses the <code>**keywords</code> syntax to accept arbitrary keyword arguments; bit <code>0x20</code> is set if the function is a generator. See <a class="reference internal" href="../library/inspect#inspect-module-co-flags"><span class="std std-ref">Code Objects Bit Flags</span></a> for details on the semantics of each flags that might be present.</p> <p>Future feature declarations (<code>from __future__ import division</code>) also use bits in <a class="reference internal" href="#codeobject.co_flags" title="codeobject.co_flags"><code>co_flags</code></a> to indicate whether a code object was compiled with a particular feature enabled: bit <code>0x2000</code> is set if the function was compiled with future division enabled; bits <code>0x10</code> and <code>0x1000</code> were used in earlier versions of Python.</p> <p>Other bits in <a class="reference internal" href="#codeobject.co_flags" title="codeobject.co_flags"><code>co_flags</code></a> are reserved for internal use.</p> <p id="index-60">If a code object represents a function, the first item in <a class="reference internal" href="#codeobject.co_consts" title="codeobject.co_consts"><code>co_consts</code></a> is the documentation string of the function, or <code>None</code> if undefined.</p> </section> <section id="methods-on-code-objects"> <h5> +<span class="section-number">3.2.13.1.2. </span>Methods on code objects</h5> <dl class="py method"> <dt class="sig sig-object py" id="codeobject.co_positions"> +<code>codeobject.co_positions()</code> </dt> <dd> +<p>Returns an iterable over the source code positions of each <a class="reference internal" href="../glossary#term-bytecode"><span class="xref std std-term">bytecode</span></a> instruction in the code object.</p> <p>The iterator returns <a class="reference internal" href="../library/stdtypes#tuple" title="tuple"><code>tuple</code></a>s containing the <code>(start_line, end_line, +start_column, end_column)</code>. The <em>i-th</em> tuple corresponds to the position of the source code that compiled to the <em>i-th</em> instruction. Column information is 0-indexed utf-8 byte offsets on the given source line.</p> <p>This positional information can be missing. A non-exhaustive lists of cases where this may happen:</p> <ul class="simple"> <li>Running the interpreter with <a class="reference internal" href="../using/cmdline#cmdoption-X"><code>-X</code></a> <code>no_debug_ranges</code>.</li> <li>Loading a pyc file compiled while using <a class="reference internal" href="../using/cmdline#cmdoption-X"><code>-X</code></a> <code>no_debug_ranges</code>.</li> <li>Position tuples corresponding to artificial instructions.</li> <li>Line and column numbers that can’t be represented due to implementation specific limitations.</li> </ul> <p>When this occurs, some or all of the tuple elements can be <a class="reference internal" href="../library/constants#None" title="None"><code>None</code></a>.</p> <div class="versionadded"> <p><span class="versionmodified added">New in version 3.11.</span></p> </div> <div class="admonition note"> <p class="admonition-title">Note</p> <p>This feature requires storing column positions in code objects which may result in a small increase of disk usage of compiled Python files or interpreter memory usage. To avoid storing the extra information and/or deactivate printing the extra traceback information, the <a class="reference internal" href="../using/cmdline#cmdoption-X"><code>-X</code></a> <code>no_debug_ranges</code> command line flag or the <span class="target" id="index-61"></span><a class="reference internal" href="../using/cmdline#envvar-PYTHONNODEBUGRANGES"><code>PYTHONNODEBUGRANGES</code></a> environment variable can be used.</p> </div> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="codeobject.co_lines"> +<code>codeobject.co_lines()</code> </dt> <dd> +<p>Returns an iterator that yields information about successive ranges of <a class="reference internal" href="../glossary#term-bytecode"><span class="xref std std-term">bytecode</span></a>s. Each item yielded is a <code>(start, end, lineno)</code> <a class="reference internal" href="../library/stdtypes#tuple" title="tuple"><code>tuple</code></a>:</p> <ul class="simple"> <li> +<code>start</code> (an <a class="reference internal" href="../library/functions#int" title="int"><code>int</code></a>) represents the offset (inclusive) of the start of the <a class="reference internal" href="../glossary#term-bytecode"><span class="xref std std-term">bytecode</span></a> range</li> <li> +<code>end</code> (an <a class="reference internal" href="../library/functions#int" title="int"><code>int</code></a>) represents the offset (inclusive) of the end of the <a class="reference internal" href="../glossary#term-bytecode"><span class="xref std std-term">bytecode</span></a> range</li> <li> +<code>lineno</code> is an <a class="reference internal" href="../library/functions#int" title="int"><code>int</code></a> representing the line number of the <a class="reference internal" href="../glossary#term-bytecode"><span class="xref std std-term">bytecode</span></a> range, or <code>None</code> if the bytecodes in the given range have no line number</li> </ul> <p>The items yielded generated will have the following properties:</p> <ul class="simple"> <li>The first range yielded will have a <code>start</code> of 0.</li> <li>The <code>(start, end)</code> ranges will be non-decreasing and consecutive. That is, for any pair of <a class="reference internal" href="../library/stdtypes#tuple" title="tuple"><code>tuple</code></a>s, the <code>start</code> of the second will be equal to the <code>end</code> of the first.</li> <li>No range will be backwards: <code>end >= start</code> for all triples.</li> <li>The <a class="reference internal" href="../library/stdtypes#tuple" title="tuple"><code>tuple</code></a> yielded will have <code>end</code> equal to the size of the <a class="reference internal" href="../glossary#term-bytecode"><span class="xref std std-term">bytecode</span></a>.</li> </ul> <p>Zero-width ranges, where <code>start == end</code>, are allowed. Zero-width ranges are used for lines that are present in the source code, but have been eliminated by the <a class="reference internal" href="../glossary#term-bytecode"><span class="xref std std-term">bytecode</span></a> compiler.</p> <div class="versionadded"> <p><span class="versionmodified added">New in version 3.10.</span></p> </div> <div class="admonition seealso"> <p class="admonition-title">See also</p> <dl class="simple"> <dt> +<span class="target" id="index-62"></span><a class="pep reference external" href="https://peps.python.org/pep-0626/"><strong>PEP 626</strong></a> - Precise line numbers for debugging and other tools.</dt> +<dd> +<p>The PEP that introduced the <code>co_lines()</code> method.</p> </dd> </dl> </div> </dd> +</dl> </section> </section> <section id="frame-objects"> <span id="id6"></span><h4> +<span class="section-number">3.2.13.2. </span>Frame objects</h4> <p id="index-63">Frame objects represent execution frames. They may occur in <a class="reference internal" href="#traceback-objects"><span class="std std-ref">traceback objects</span></a>, and are also passed to registered trace functions.</p> <section id="index-64"> <span id="id7"></span><h5> +<span class="section-number">3.2.13.2.1. </span>Special read-only attributes</h5> <table class="docutils align-default"> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="frame.f_back"> +<code>frame.f_back</code> </dt> <dd></dd> +</dl> </td> <td><p>Points to the previous stack frame (towards the caller), or <code>None</code> if this is the bottom stack frame</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="frame.f_code"> +<code>frame.f_code</code> </dt> <dd></dd> +</dl> </td> <td><p>The <a class="reference internal" href="#code-objects"><span class="std std-ref">code object</span></a> being executed in this frame. Accessing this attribute raises an <a class="reference internal" href="../library/sys#auditing"><span class="std std-ref">auditing event</span></a> <code>object.__getattr__</code> with arguments <code>obj</code> and <code>"f_code"</code>.</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="frame.f_locals"> +<code>frame.f_locals</code> </dt> <dd></dd> +</dl> </td> <td><p>The dictionary used by the frame to look up <a class="reference internal" href="executionmodel#naming"><span class="std std-ref">local variables</span></a></p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="frame.f_globals"> +<code>frame.f_globals</code> </dt> <dd></dd> +</dl> </td> <td><p>The dictionary used by the frame to look up <a class="reference internal" href="executionmodel#naming"><span class="std std-ref">global variables</span></a></p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="frame.f_builtins"> +<code>frame.f_builtins</code> </dt> <dd></dd> +</dl> </td> <td><p>The dictionary used by the frame to look up <a class="reference internal" href="executionmodel#naming"><span class="std std-ref">built-in (intrinsic) names</span></a></p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="frame.f_lasti"> +<code>frame.f_lasti</code> </dt> <dd></dd> +</dl> </td> <td><p>The “precise instruction” of the frame object (this is an index into the <a class="reference internal" href="../glossary#term-bytecode"><span class="xref std std-term">bytecode</span></a> string of the <a class="reference internal" href="#code-objects"><span class="std std-ref">code object</span></a>)</p></td> </tr> </table> </section> <section id="index-65"> <span id="id8"></span><h5> +<span class="section-number">3.2.13.2.2. </span>Special writable attributes</h5> <table class="docutils align-default"> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="frame.f_trace"> +<code>frame.f_trace</code> </dt> <dd></dd> +</dl> </td> <td><p>If not <code>None</code>, this is a function called for various events during code execution (this is used by debuggers). Normally an event is triggered for each new source line (see <a class="reference internal" href="#frame.f_trace_lines" title="frame.f_trace_lines"><code>f_trace_lines</code></a>).</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="frame.f_trace_lines"> +<code>frame.f_trace_lines</code> </dt> <dd></dd> +</dl> </td> <td><p>Set this attribute to <a class="reference internal" href="../library/constants#False" title="False"><code>False</code></a> to disable triggering a tracing event for each source line.</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="frame.f_trace_opcodes"> +<code>frame.f_trace_opcodes</code> </dt> <dd></dd> +</dl> </td> <td><p>Set this attribute to <a class="reference internal" href="../library/constants#True" title="True"><code>True</code></a> to allow per-opcode events to be requested. Note that this may lead to undefined interpreter behaviour if exceptions raised by the trace function escape to the function being traced.</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="frame.f_lineno"> +<code>frame.f_lineno</code> </dt> <dd></dd> +</dl> </td> <td><p>The current line number of the frame – writing to this from within a trace function jumps to the given line (only for the bottom-most frame). A debugger can implement a Jump command (aka Set Next Statement) by writing to this attribute.</p></td> </tr> </table> </section> <section id="frame-object-methods"> <h5> +<span class="section-number">3.2.13.2.3. </span>Frame object methods</h5> <p>Frame objects support one method:</p> <dl class="py method"> <dt class="sig sig-object py" id="frame.clear"> +<code>frame.clear()</code> </dt> <dd> +<p>This method clears all references to <a class="reference internal" href="executionmodel#naming"><span class="std std-ref">local variables</span></a> held by the frame. Also, if the frame belonged to a <a class="reference internal" href="../glossary#term-generator"><span class="xref std std-term">generator</span></a>, the generator is finalized. This helps break reference cycles involving frame objects (for example when catching an <a class="reference internal" href="../library/exceptions#bltin-exceptions"><span class="std std-ref">exception</span></a> and storing its <a class="reference internal" href="#traceback-objects"><span class="std std-ref">traceback</span></a> for later use).</p> <p><a class="reference internal" href="../library/exceptions#RuntimeError" title="RuntimeError"><code>RuntimeError</code></a> is raised if the frame is currently executing.</p> <div class="versionadded"> <p><span class="versionmodified added">New in version 3.4.</span></p> </div> </dd> +</dl> </section> </section> <section id="traceback-objects"> <span id="id9"></span><h4> +<span class="section-number">3.2.13.3. </span>Traceback objects</h4> <p id="index-66">Traceback objects represent the stack trace of an <a class="reference internal" href="../tutorial/errors#tut-errors"><span class="std std-ref">exception</span></a>. A traceback object is implicitly created when an exception occurs, and may also be explicitly created by calling <a class="reference internal" href="../library/types#types.TracebackType" title="types.TracebackType"><code>types.TracebackType</code></a>.</p> <div class="versionchanged"> <p><span class="versionmodified changed">Changed in version 3.7: </span>Traceback objects can now be explicitly instantiated from Python code.</p> </div> <p>For implicitly created tracebacks, when the search for an exception handler unwinds the execution stack, at each unwound level a traceback object is inserted in front of the current traceback. When an exception handler is entered, the stack trace is made available to the program. (See section <a class="reference internal" href="compound_stmts#try"><span class="std std-ref">The try statement</span></a>.) It is accessible as the third item of the tuple returned by <a class="reference internal" href="../library/sys#sys.exc_info" title="sys.exc_info"><code>sys.exc_info()</code></a>, and as the <a class="reference internal" href="../library/exceptions#BaseException.__traceback__" title="BaseException.__traceback__"><code>__traceback__</code></a> attribute of the caught exception.</p> <p>When the program contains no suitable handler, the stack trace is written (nicely formatted) to the standard error stream; if the interpreter is interactive, it is also made available to the user as <a class="reference internal" href="../library/sys#sys.last_traceback" title="sys.last_traceback"><code>sys.last_traceback</code></a>.</p> <p>For explicitly created tracebacks, it is up to the creator of the traceback to determine how the <a class="reference internal" href="#traceback.tb_next" title="traceback.tb_next"><code>tb_next</code></a> attributes should be linked to form a full stack trace.</p> <p id="index-67">Special read-only attributes:</p> <table class="docutils align-default"> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="traceback.tb_frame"> +<code>traceback.tb_frame</code> </dt> <dd></dd> +</dl> </td> <td> +<p>Points to the execution <a class="reference internal" href="#frame-objects"><span class="std std-ref">frame</span></a> of the current level.</p> <p>Accessing this attribute raises an <a class="reference internal" href="../library/sys#auditing"><span class="std std-ref">auditing event</span></a> <code>object.__getattr__</code> with arguments <code>obj</code> and <code>"tb_frame"</code>.</p> </td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="traceback.tb_lineno"> +<code>traceback.tb_lineno</code> </dt> <dd></dd> +</dl> </td> <td><p>Gives the line number where the exception occurred</p></td> </tr> <tr> +<td> +<dl class="py attribute"> <dt class="sig sig-object py" id="traceback.tb_lasti"> +<code>traceback.tb_lasti</code> </dt> <dd></dd> +</dl> </td> <td><p>Indicates the “precise instruction”.</p></td> </tr> </table> <p>The line number and last instruction in the traceback may differ from the line number of its <a class="reference internal" href="#frame-objects"><span class="std std-ref">frame object</span></a> if the exception occurred in a <a class="reference internal" href="compound_stmts#try"><code>try</code></a> statement with no matching except clause or with a <a class="reference internal" href="compound_stmts#finally"><code>finally</code></a> clause.</p> <span class="target" id="index-68"></span><dl class="py attribute"> <dt class="sig sig-object py" id="traceback.tb_next"> +<code>traceback.tb_next</code> </dt> <dd> +<p>The special writable attribute <code>tb_next</code> is the next level in the stack trace (towards the frame where the exception occurred), or <code>None</code> if there is no next level.</p> <div class="versionchanged"> <p><span class="versionmodified changed">Changed in version 3.7: </span>This attribute is now writable</p> </div> </dd> +</dl> </section> <section id="slice-objects"> <h4> +<span class="section-number">3.2.13.4. </span>Slice objects</h4> <p id="index-69">Slice objects are used to represent slices for <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><code>__getitem__()</code></a> methods. They are also created by the built-in <a class="reference internal" href="../library/functions#slice" title="slice"><code>slice()</code></a> function.</p> <p id="index-70">Special read-only attributes: <a class="reference internal" href="../library/functions#slice.start" title="slice.start"><code>start</code></a> is the lower bound; <a class="reference internal" href="../library/functions#slice.stop" title="slice.stop"><code>stop</code></a> is the upper bound; <a class="reference internal" href="../library/functions#slice.step" title="slice.step"><code>step</code></a> is the step value; each is <code>None</code> if omitted. These attributes can have any type.</p> <p>Slice objects support one method:</p> <dl class="py method"> <dt class="sig sig-object py" id="slice.indices"> +<code>slice.indices(self, length)</code> </dt> <dd> +<p>This method takes a single integer argument <em>length</em> and computes information about the slice that the slice object would describe if applied to a sequence of <em>length</em> items. It returns a tuple of three integers; respectively these are the <em>start</em> and <em>stop</em> indices and the <em>step</em> or stride length of the slice. Missing or out-of-bounds indices are handled in a manner consistent with regular slices.</p> </dd> +</dl> </section> <section id="static-method-objects"> <h4> +<span class="section-number">3.2.13.5. </span>Static method objects</h4> <p>Static method objects provide a way of defeating the transformation of function objects to method objects described above. A static method object is a wrapper around any other object, usually a user-defined method object. When a static method object is retrieved from a class or a class instance, the object actually returned is the wrapped object, which is not subject to any further transformation. Static method objects are also callable. Static method objects are created by the built-in <a class="reference internal" href="../library/functions#staticmethod" title="staticmethod"><code>staticmethod()</code></a> constructor.</p> </section> <section id="class-method-objects"> <h4> +<span class="section-number">3.2.13.6. </span>Class method objects</h4> <p>A class method object, like a static method object, is a wrapper around another object that alters the way in which that object is retrieved from classes and class instances. The behaviour of class method objects upon such retrieval is described above, under “User-defined methods”. Class method objects are created by the built-in <a class="reference internal" href="../library/functions#classmethod" title="classmethod"><code>classmethod()</code></a> constructor.</p> </section> </section> </section> <section id="special-method-names"> <span id="specialnames"></span><h2> +<span class="section-number">3.3. </span>Special method names</h2> <p id="index-71">A class can implement certain operations that are invoked by special syntax (such as arithmetic operations or subscripting and slicing) by defining methods with special names. This is Python’s approach to <em class="dfn">operator overloading</em>, allowing classes to define their own behavior with respect to language operators. For instance, if a class defines a method named <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><code>__getitem__()</code></a>, and <code>x</code> is an instance of this class, then <code>x[i]</code> is roughly equivalent to <code>type(x).__getitem__(x, i)</code>. Except where mentioned, attempts to execute an operation raise an exception when no appropriate method is defined (typically <a class="reference internal" href="../library/exceptions#AttributeError" title="AttributeError"><code>AttributeError</code></a> or <a class="reference internal" href="../library/exceptions#TypeError" title="TypeError"><code>TypeError</code></a>).</p> <p>Setting a special method to <code>None</code> indicates that the corresponding operation is not available. For example, if a class sets <a class="reference internal" href="#object.__iter__" title="object.__iter__"><code>__iter__()</code></a> to <code>None</code>, the class is not iterable, so calling <a class="reference internal" href="../library/functions#iter" title="iter"><code>iter()</code></a> on its instances will raise a <a class="reference internal" href="../library/exceptions#TypeError" title="TypeError"><code>TypeError</code></a> (without falling back to <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><code>__getitem__()</code></a>). <a class="footnote-reference brackets" href="#id18" id="id10">2</a></p> <p>When implementing a class that emulates any built-in type, it is important that the emulation only be implemented to the degree that it makes sense for the object being modelled. For example, some sequences may work well with retrieval of individual elements, but extracting a slice may not make sense. (One example of this is the <code>NodeList</code> interface in the W3C’s Document Object Model.)</p> <section id="basic-customization"> <span id="customization"></span><h3> +<span class="section-number">3.3.1. </span>Basic customization</h3> <dl class="py method"> <dt class="sig sig-object py" id="object.__new__"> +<code>object.__new__(cls[, ...])</code> </dt> <dd> +<p id="index-72">Called to create a new instance of class <em>cls</em>. <a class="reference internal" href="#object.__new__" title="object.__new__"><code>__new__()</code></a> is a static method (special-cased so you need not declare it as such) that takes the class of which an instance was requested as its first argument. The remaining arguments are those passed to the object constructor expression (the call to the class). The return value of <a class="reference internal" href="#object.__new__" title="object.__new__"><code>__new__()</code></a> should be the new object instance (usually an instance of <em>cls</em>).</p> <p>Typical implementations create a new instance of the class by invoking the superclass’s <a class="reference internal" href="#object.__new__" title="object.__new__"><code>__new__()</code></a> method using <code>super().__new__(cls[, ...])</code> with appropriate arguments and then modifying the newly created instance as necessary before returning it.</p> <p>If <a class="reference internal" href="#object.__new__" title="object.__new__"><code>__new__()</code></a> is invoked during object construction and it returns an instance of <em>cls</em>, then the new instance’s <a class="reference internal" href="#object.__init__" title="object.__init__"><code>__init__()</code></a> method will be invoked like <code>__init__(self[, ...])</code>, where <em>self</em> is the new instance and the remaining arguments are the same as were passed to the object constructor.</p> <p>If <a class="reference internal" href="#object.__new__" title="object.__new__"><code>__new__()</code></a> does not return an instance of <em>cls</em>, then the new instance’s <a class="reference internal" href="#object.__init__" title="object.__init__"><code>__init__()</code></a> method will not be invoked.</p> <p><a class="reference internal" href="#object.__new__" title="object.__new__"><code>__new__()</code></a> is intended mainly to allow subclasses of immutable types (like int, str, or tuple) to customize instance creation. It is also commonly overridden in custom metaclasses in order to customize class creation.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__init__"> +<code>object.__init__(self[, ...])</code> </dt> <dd> +<p id="index-73">Called after the instance has been created (by <a class="reference internal" href="#object.__new__" title="object.__new__"><code>__new__()</code></a>), but before it is returned to the caller. The arguments are those passed to the class constructor expression. If a base class has an <a class="reference internal" href="#object.__init__" title="object.__init__"><code>__init__()</code></a> method, the derived class’s <a class="reference internal" href="#object.__init__" title="object.__init__"><code>__init__()</code></a> method, if any, must explicitly call it to ensure proper initialization of the base class part of the instance; for example: <code>super().__init__([args...])</code>.</p> <p>Because <a class="reference internal" href="#object.__new__" title="object.__new__"><code>__new__()</code></a> and <a class="reference internal" href="#object.__init__" title="object.__init__"><code>__init__()</code></a> work together in constructing objects (<a class="reference internal" href="#object.__new__" title="object.__new__"><code>__new__()</code></a> to create it, and <a class="reference internal" href="#object.__init__" title="object.__init__"><code>__init__()</code></a> to customize it), no non-<code>None</code> value may be returned by <a class="reference internal" href="#object.__init__" title="object.__init__"><code>__init__()</code></a>; doing so will cause a <a class="reference internal" href="../library/exceptions#TypeError" title="TypeError"><code>TypeError</code></a> to be raised at runtime.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__del__"> +<code>object.__del__(self)</code> </dt> <dd> +<p id="index-74">Called when the instance is about to be destroyed. This is also called a finalizer or (improperly) a destructor. If a base class has a <a class="reference internal" href="#object.__del__" title="object.__del__"><code>__del__()</code></a> method, the derived class’s <a class="reference internal" href="#object.__del__" title="object.__del__"><code>__del__()</code></a> method, if any, must explicitly call it to ensure proper deletion of the base class part of the instance.</p> <p>It is possible (though not recommended!) for the <a class="reference internal" href="#object.__del__" title="object.__del__"><code>__del__()</code></a> method to postpone destruction of the instance by creating a new reference to it. This is called object <em>resurrection</em>. It is implementation-dependent whether <a class="reference internal" href="#object.__del__" title="object.__del__"><code>__del__()</code></a> is called a second time when a resurrected object is about to be destroyed; the current <a class="reference internal" href="../glossary#term-CPython"><span class="xref std std-term">CPython</span></a> implementation only calls it once.</p> <p>It is not guaranteed that <a class="reference internal" href="#object.__del__" title="object.__del__"><code>__del__()</code></a> methods are called for objects that still exist when the interpreter exits.</p> <div class="admonition note"> <p class="admonition-title">Note</p> <p><code>del x</code> doesn’t directly call <code>x.__del__()</code> — the former decrements the reference count for <code>x</code> by one, and the latter is only called when <code>x</code>’s reference count reaches zero.</p> </div> <div class="impl-detail compound"> <p class="compound-first"><strong>CPython implementation detail:</strong> It is possible for a reference cycle to prevent the reference count of an object from going to zero. In this case, the cycle will be later detected and deleted by the <a class="reference internal" href="../glossary#term-garbage-collection"><span class="xref std std-term">cyclic garbage collector</span></a>. A common cause of reference cycles is when an exception has been caught in a local variable. The frame’s locals then reference the exception, which references its own traceback, which references the locals of all frames caught in the traceback.</p> <div class="compound-last admonition seealso"> <p class="admonition-title">See also</p> <p>Documentation for the <a class="reference internal" href="../library/gc#module-gc" title="gc: Interface to the cycle-detecting garbage collector."><code>gc</code></a> module.</p> </div> </div> <div class="admonition warning"> <p class="admonition-title">Warning</p> <p>Due to the precarious circumstances under which <a class="reference internal" href="#object.__del__" title="object.__del__"><code>__del__()</code></a> methods are invoked, exceptions that occur during their execution are ignored, and a warning is printed to <code>sys.stderr</code> instead. In particular:</p> <ul class="simple"> <li> +<a class="reference internal" href="#object.__del__" title="object.__del__"><code>__del__()</code></a> can be invoked when arbitrary code is being executed, including from any arbitrary thread. If <a class="reference internal" href="#object.__del__" title="object.__del__"><code>__del__()</code></a> needs to take a lock or invoke any other blocking resource, it may deadlock as the resource may already be taken by the code that gets interrupted to execute <a class="reference internal" href="#object.__del__" title="object.__del__"><code>__del__()</code></a>.</li> <li> +<a class="reference internal" href="#object.__del__" title="object.__del__"><code>__del__()</code></a> can be executed during interpreter shutdown. As a consequence, the global variables it needs to access (including other modules) may already have been deleted or set to <code>None</code>. Python guarantees that globals whose name begins with a single underscore are deleted from their module before other globals are deleted; if no other references to such globals exist, this may help in assuring that imported modules are still available at the time when the <a class="reference internal" href="#object.__del__" title="object.__del__"><code>__del__()</code></a> method is called.</li> </ul> </div> <span class="target" id="index-75"></span> +</dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__repr__"> +<code>object.__repr__(self)</code> </dt> <dd> +<p>Called by the <a class="reference internal" href="../library/functions#repr" title="repr"><code>repr()</code></a> built-in function to compute the “official” string representation of an object. If at all possible, this should look like a valid Python expression that could be used to recreate an object with the same value (given an appropriate environment). If this is not possible, a string of the form <code><...some useful description...></code> should be returned. The return value must be a string object. If a class defines <a class="reference internal" href="#object.__repr__" title="object.__repr__"><code>__repr__()</code></a> but not <a class="reference internal" href="#object.__str__" title="object.__str__"><code>__str__()</code></a>, then <a class="reference internal" href="#object.__repr__" title="object.__repr__"><code>__repr__()</code></a> is also used when an “informal” string representation of instances of that class is required.</p> <p>This is typically used for debugging, so it is important that the representation is information-rich and unambiguous.</p> <span class="target" id="index-76"></span> +</dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__str__"> +<code>object.__str__(self)</code> </dt> <dd> +<p>Called by <a class="reference internal" href="../library/stdtypes#str" title="str"><code>str(object)</code></a> and the built-in functions <a class="reference internal" href="../library/functions#format" title="format"><code>format()</code></a> and <a class="reference internal" href="../library/functions#print" title="print"><code>print()</code></a> to compute the “informal” or nicely printable string representation of an object. The return value must be a <a class="reference internal" href="../library/stdtypes#textseq"><span class="std std-ref">string</span></a> object.</p> <p>This method differs from <a class="reference internal" href="#object.__repr__" title="object.__repr__"><code>object.__repr__()</code></a> in that there is no expectation that <a class="reference internal" href="#object.__str__" title="object.__str__"><code>__str__()</code></a> return a valid Python expression: a more convenient or concise representation can be used.</p> <p>The default implementation defined by the built-in type <a class="reference internal" href="../library/functions#object" title="object"><code>object</code></a> calls <a class="reference internal" href="#object.__repr__" title="object.__repr__"><code>object.__repr__()</code></a>.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__bytes__"> +<code>object.__bytes__(self)</code> </dt> <dd> +<p id="index-77">Called by <a class="reference internal" href="../library/functions#func-bytes"><span class="std std-ref">bytes</span></a> to compute a byte-string representation of an object. This should return a <a class="reference internal" href="../library/stdtypes#bytes" title="bytes"><code>bytes</code></a> object.</p> <span class="target" id="index-78"></span> +</dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__format__"> +<code>object.__format__(self, format_spec)</code> </dt> <dd> +<p>Called by the <a class="reference internal" href="../library/functions#format" title="format"><code>format()</code></a> built-in function, and by extension, evaluation of <a class="reference internal" href="lexical_analysis#f-strings"><span class="std std-ref">formatted string literals</span></a> and the <a class="reference internal" href="../library/stdtypes#str.format" title="str.format"><code>str.format()</code></a> method, to produce a “formatted” string representation of an object. The <em>format_spec</em> argument is a string that contains a description of the formatting options desired. The interpretation of the <em>format_spec</em> argument is up to the type implementing <a class="reference internal" href="#object.__format__" title="object.__format__"><code>__format__()</code></a>, however most classes will either delegate formatting to one of the built-in types, or use a similar formatting option syntax.</p> <p>See <a class="reference internal" href="../library/string#formatspec"><span class="std std-ref">Format Specification Mini-Language</span></a> for a description of the standard formatting syntax.</p> <p>The return value must be a string object.</p> <div class="versionchanged"> <p><span class="versionmodified changed">Changed in version 3.4: </span>The __format__ method of <code>object</code> itself raises a <a class="reference internal" href="../library/exceptions#TypeError" title="TypeError"><code>TypeError</code></a> if passed any non-empty string.</p> </div> <div class="versionchanged"> <p><span class="versionmodified changed">Changed in version 3.7: </span><code>object.__format__(x, '')</code> is now equivalent to <code>str(x)</code> rather than <code>format(str(x), '')</code>.</p> </div> </dd> +</dl> <span class="target" id="richcmpfuncs"></span><dl class="py method"> <dt class="sig sig-object py" id="object.__lt__"> +<code>object.__lt__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__le__"> +<code>object.__le__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__eq__"> +<code>object.__eq__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__ne__"> +<code>object.__ne__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__gt__"> +<code>object.__gt__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__ge__"> +<code>object.__ge__(self, other)</code> </dt> <dd> +<p id="index-79">These are the so-called “rich comparison” methods. The correspondence between operator symbols and method names is as follows: <code>x<y</code> calls <code>x.__lt__(y)</code>, <code>x<=y</code> calls <code>x.__le__(y)</code>, <code>x==y</code> calls <code>x.__eq__(y)</code>, <code>x!=y</code> calls <code>x.__ne__(y)</code>, <code>x>y</code> calls <code>x.__gt__(y)</code>, and <code>x>=y</code> calls <code>x.__ge__(y)</code>.</p> <p>A rich comparison method may return the singleton <code>NotImplemented</code> if it does not implement the operation for a given pair of arguments. By convention, <code>False</code> and <code>True</code> are returned for a successful comparison. However, these methods can return any value, so if the comparison operator is used in a Boolean context (e.g., in the condition of an <code>if</code> statement), Python will call <a class="reference internal" href="../library/functions#bool" title="bool"><code>bool()</code></a> on the value to determine if the result is true or false.</p> <p>By default, <code>object</code> implements <a class="reference internal" href="#object.__eq__" title="object.__eq__"><code>__eq__()</code></a> by using <code>is</code>, returning <code>NotImplemented</code> in the case of a false comparison: <code>True if x is y else NotImplemented</code>. For <a class="reference internal" href="#object.__ne__" title="object.__ne__"><code>__ne__()</code></a>, by default it delegates to <a class="reference internal" href="#object.__eq__" title="object.__eq__"><code>__eq__()</code></a> and inverts the result unless it is <code>NotImplemented</code>. There are no other implied relationships among the comparison operators or default implementations; for example, the truth of <code>(x<y or x==y)</code> does not imply <code>x<=y</code>. To automatically generate ordering operations from a single root operation, see <a class="reference internal" href="../library/functools#functools.total_ordering" title="functools.total_ordering"><code>functools.total_ordering()</code></a>.</p> <p>See the paragraph on <a class="reference internal" href="#object.__hash__" title="object.__hash__"><code>__hash__()</code></a> for some important notes on creating <a class="reference internal" href="../glossary#term-hashable"><span class="xref std std-term">hashable</span></a> objects which support custom comparison operations and are usable as dictionary keys.</p> <p>There are no swapped-argument versions of these methods (to be used when the left argument does not support the operation but the right argument does); rather, <a class="reference internal" href="#object.__lt__" title="object.__lt__"><code>__lt__()</code></a> and <a class="reference internal" href="#object.__gt__" title="object.__gt__"><code>__gt__()</code></a> are each other’s reflection, <a class="reference internal" href="#object.__le__" title="object.__le__"><code>__le__()</code></a> and <a class="reference internal" href="#object.__ge__" title="object.__ge__"><code>__ge__()</code></a> are each other’s reflection, and <a class="reference internal" href="#object.__eq__" title="object.__eq__"><code>__eq__()</code></a> and <a class="reference internal" href="#object.__ne__" title="object.__ne__"><code>__ne__()</code></a> are their own reflection. If the operands are of different types, and right operand’s type is a direct or indirect subclass of the left operand’s type, the reflected method of the right operand has priority, otherwise the left operand’s method has priority. Virtual subclassing is not considered.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__hash__"> +<code>object.__hash__(self)</code> </dt> <dd> +<p id="index-80">Called by built-in function <a class="reference internal" href="../library/functions#hash" title="hash"><code>hash()</code></a> and for operations on members of hashed collections including <a class="reference internal" href="../library/stdtypes#set" title="set"><code>set</code></a>, <a class="reference internal" href="../library/stdtypes#frozenset" title="frozenset"><code>frozenset</code></a>, and <a class="reference internal" href="../library/stdtypes#dict" title="dict"><code>dict</code></a>. The <code>__hash__()</code> method should return an integer. The only required property is that objects which compare equal have the same hash value; it is advised to mix together the hash values of the components of the object that also play a part in comparison of objects by packing them into a tuple and hashing the tuple. Example:</p> <pre data-language="python">def __hash__(self): + return hash((self.name, self.nick, self.color)) +</pre> <div class="admonition note"> <p class="admonition-title">Note</p> <p><a class="reference internal" href="../library/functions#hash" title="hash"><code>hash()</code></a> truncates the value returned from an object’s custom <a class="reference internal" href="#object.__hash__" title="object.__hash__"><code>__hash__()</code></a> method to the size of a <a class="reference internal" href="../c-api/intro#c.Py_ssize_t" title="Py_ssize_t"><code>Py_ssize_t</code></a>. This is typically 8 bytes on 64-bit builds and 4 bytes on 32-bit builds. If an object’s <a class="reference internal" href="#object.__hash__" title="object.__hash__"><code>__hash__()</code></a> must interoperate on builds of different bit sizes, be sure to check the width on all supported builds. An easy way to do this is with <code>python -c "import sys; print(sys.hash_info.width)"</code>.</p> </div> <p>If a class does not define an <a class="reference internal" href="#object.__eq__" title="object.__eq__"><code>__eq__()</code></a> method it should not define a <a class="reference internal" href="#object.__hash__" title="object.__hash__"><code>__hash__()</code></a> operation either; if it defines <a class="reference internal" href="#object.__eq__" title="object.__eq__"><code>__eq__()</code></a> but not <a class="reference internal" href="#object.__hash__" title="object.__hash__"><code>__hash__()</code></a>, its instances will not be usable as items in hashable collections. If a class defines mutable objects and implements an <a class="reference internal" href="#object.__eq__" title="object.__eq__"><code>__eq__()</code></a> method, it should not implement <a class="reference internal" href="#object.__hash__" title="object.__hash__"><code>__hash__()</code></a>, since the implementation of <a class="reference internal" href="../glossary#term-hashable"><span class="xref std std-term">hashable</span></a> collections requires that a key’s hash value is immutable (if the object’s hash value changes, it will be in the wrong hash bucket).</p> <p>User-defined classes have <a class="reference internal" href="#object.__eq__" title="object.__eq__"><code>__eq__()</code></a> and <a class="reference internal" href="#object.__hash__" title="object.__hash__"><code>__hash__()</code></a> methods by default; with them, all objects compare unequal (except with themselves) and <code>x.__hash__()</code> returns an appropriate value such that <code>x == y</code> implies both that <code>x is y</code> and <code>hash(x) == hash(y)</code>.</p> <p>A class that overrides <a class="reference internal" href="#object.__eq__" title="object.__eq__"><code>__eq__()</code></a> and does not define <a class="reference internal" href="#object.__hash__" title="object.__hash__"><code>__hash__()</code></a> will have its <a class="reference internal" href="#object.__hash__" title="object.__hash__"><code>__hash__()</code></a> implicitly set to <code>None</code>. When the <a class="reference internal" href="#object.__hash__" title="object.__hash__"><code>__hash__()</code></a> method of a class is <code>None</code>, instances of the class will raise an appropriate <a class="reference internal" href="../library/exceptions#TypeError" title="TypeError"><code>TypeError</code></a> when a program attempts to retrieve their hash value, and will also be correctly identified as unhashable when checking <code>isinstance(obj, collections.abc.Hashable)</code>.</p> <p>If a class that overrides <a class="reference internal" href="#object.__eq__" title="object.__eq__"><code>__eq__()</code></a> needs to retain the implementation of <a class="reference internal" href="#object.__hash__" title="object.__hash__"><code>__hash__()</code></a> from a parent class, the interpreter must be told this explicitly by setting <code>__hash__ = <ParentClass>.__hash__</code>.</p> <p>If a class that does not override <a class="reference internal" href="#object.__eq__" title="object.__eq__"><code>__eq__()</code></a> wishes to suppress hash support, it should include <code>__hash__ = None</code> in the class definition. A class which defines its own <a class="reference internal" href="#object.__hash__" title="object.__hash__"><code>__hash__()</code></a> that explicitly raises a <a class="reference internal" href="../library/exceptions#TypeError" title="TypeError"><code>TypeError</code></a> would be incorrectly identified as hashable by an <code>isinstance(obj, collections.abc.Hashable)</code> call.</p> <div class="admonition note"> <p class="admonition-title">Note</p> <p>By default, the <a class="reference internal" href="#object.__hash__" title="object.__hash__"><code>__hash__()</code></a> values of str and bytes objects are “salted” with an unpredictable random value. Although they remain constant within an individual Python process, they are not predictable between repeated invocations of Python.</p> <p>This is intended to provide protection against a denial-of-service caused by carefully chosen inputs that exploit the worst case performance of a dict insertion, O(n<sup>2</sup>) complexity. See <a class="reference external" href="http://ocert.org/advisories/ocert-2011-003.html">http://ocert.org/advisories/ocert-2011-003.html</a> for details.</p> <p>Changing hash values affects the iteration order of sets. Python has never made guarantees about this ordering (and it typically varies between 32-bit and 64-bit builds).</p> <p>See also <span class="target" id="index-81"></span><a class="reference internal" href="../using/cmdline#envvar-PYTHONHASHSEED"><code>PYTHONHASHSEED</code></a>.</p> </div> <div class="versionchanged"> <p><span class="versionmodified changed">Changed in version 3.3: </span>Hash randomization is enabled by default.</p> </div> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__bool__"> +<code>object.__bool__(self)</code> </dt> <dd> +<p id="index-82">Called to implement truth value testing and the built-in operation <code>bool()</code>; should return <code>False</code> or <code>True</code>. When this method is not defined, <a class="reference internal" href="#object.__len__" title="object.__len__"><code>__len__()</code></a> is called, if it is defined, and the object is considered true if its result is nonzero. If a class defines neither <code>__len__()</code> nor <code>__bool__()</code>, all its instances are considered true.</p> </dd> +</dl> </section> <section id="customizing-attribute-access"> <span id="attribute-access"></span><h3> +<span class="section-number">3.3.2. </span>Customizing attribute access</h3> <p>The following methods can be defined to customize the meaning of attribute access (use of, assignment to, or deletion of <code>x.name</code>) for class instances.</p> <dl class="py method"> <dt class="sig sig-object py" id="object.__getattr__"> +<code>object.__getattr__(self, name)</code> </dt> <dd> +<p>Called when the default attribute access fails with an <a class="reference internal" href="../library/exceptions#AttributeError" title="AttributeError"><code>AttributeError</code></a> (either <a class="reference internal" href="#object.__getattribute__" title="object.__getattribute__"><code>__getattribute__()</code></a> raises an <a class="reference internal" href="../library/exceptions#AttributeError" title="AttributeError"><code>AttributeError</code></a> because <em>name</em> is not an instance attribute or an attribute in the class tree for <code>self</code>; or <a class="reference internal" href="#object.__get__" title="object.__get__"><code>__get__()</code></a> of a <em>name</em> property raises <a class="reference internal" href="../library/exceptions#AttributeError" title="AttributeError"><code>AttributeError</code></a>). This method should either return the (computed) attribute value or raise an <a class="reference internal" href="../library/exceptions#AttributeError" title="AttributeError"><code>AttributeError</code></a> exception.</p> <p>Note that if the attribute is found through the normal mechanism, <a class="reference internal" href="#object.__getattr__" title="object.__getattr__"><code>__getattr__()</code></a> is not called. (This is an intentional asymmetry between <a class="reference internal" href="#object.__getattr__" title="object.__getattr__"><code>__getattr__()</code></a> and <a class="reference internal" href="#object.__setattr__" title="object.__setattr__"><code>__setattr__()</code></a>.) This is done both for efficiency reasons and because otherwise <a class="reference internal" href="#object.__getattr__" title="object.__getattr__"><code>__getattr__()</code></a> would have no way to access other attributes of the instance. Note that at least for instance variables, you can fake total control by not inserting any values in the instance attribute dictionary (but instead inserting them in another object). See the <a class="reference internal" href="#object.__getattribute__" title="object.__getattribute__"><code>__getattribute__()</code></a> method below for a way to actually get total control over attribute access.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__getattribute__"> +<code>object.__getattribute__(self, name)</code> </dt> <dd> +<p>Called unconditionally to implement attribute accesses for instances of the class. If the class also defines <a class="reference internal" href="#object.__getattr__" title="object.__getattr__"><code>__getattr__()</code></a>, the latter will not be called unless <a class="reference internal" href="#object.__getattribute__" title="object.__getattribute__"><code>__getattribute__()</code></a> either calls it explicitly or raises an <a class="reference internal" href="../library/exceptions#AttributeError" title="AttributeError"><code>AttributeError</code></a>. This method should return the (computed) attribute value or raise an <a class="reference internal" href="../library/exceptions#AttributeError" title="AttributeError"><code>AttributeError</code></a> exception. In order to avoid infinite recursion in this method, its implementation should always call the base class method with the same name to access any attributes it needs, for example, <code>object.__getattribute__(self, name)</code>.</p> <div class="admonition note"> <p class="admonition-title">Note</p> <p>This method may still be bypassed when looking up special methods as the result of implicit invocation via language syntax or <a class="reference internal" href="#builtin-functions"><span class="std std-ref">built-in functions</span></a>. See <a class="reference internal" href="#special-lookup"><span class="std std-ref">Special method lookup</span></a>.</p> </div> <p class="audit-hook"></p> +<p>For certain sensitive attribute accesses, raises an <a class="reference internal" href="../library/sys#auditing"><span class="std std-ref">auditing event</span></a> <code>object.__getattr__</code> with arguments <code>obj</code> and <code>name</code>.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__setattr__"> +<code>object.__setattr__(self, name, value)</code> </dt> <dd> +<p>Called when an attribute assignment is attempted. This is called instead of the normal mechanism (i.e. store the value in the instance dictionary). <em>name</em> is the attribute name, <em>value</em> is the value to be assigned to it.</p> <p>If <a class="reference internal" href="#object.__setattr__" title="object.__setattr__"><code>__setattr__()</code></a> wants to assign to an instance attribute, it should call the base class method with the same name, for example, <code>object.__setattr__(self, name, value)</code>.</p> <p class="audit-hook"></p> +<p>For certain sensitive attribute assignments, raises an <a class="reference internal" href="../library/sys#auditing"><span class="std std-ref">auditing event</span></a> <code>object.__setattr__</code> with arguments <code>obj</code>, <code>name</code>, <code>value</code>.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__delattr__"> +<code>object.__delattr__(self, name)</code> </dt> <dd> +<p>Like <a class="reference internal" href="#object.__setattr__" title="object.__setattr__"><code>__setattr__()</code></a> but for attribute deletion instead of assignment. This should only be implemented if <code>del obj.name</code> is meaningful for the object.</p> <p class="audit-hook"></p> +<p>For certain sensitive attribute deletions, raises an <a class="reference internal" href="../library/sys#auditing"><span class="std std-ref">auditing event</span></a> <code>object.__delattr__</code> with arguments <code>obj</code> and <code>name</code>.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__dir__"> +<code>object.__dir__(self)</code> </dt> <dd> +<p>Called when <a class="reference internal" href="../library/functions#dir" title="dir"><code>dir()</code></a> is called on the object. A sequence must be returned. <a class="reference internal" href="../library/functions#dir" title="dir"><code>dir()</code></a> converts the returned sequence to a list and sorts it.</p> </dd> +</dl> <section id="customizing-module-attribute-access"> <h4> +<span class="section-number">3.3.2.1. </span>Customizing module attribute access</h4> <p id="index-83">Special names <code>__getattr__</code> and <code>__dir__</code> can be also used to customize access to module attributes. The <code>__getattr__</code> function at the module level should accept one argument which is the name of an attribute and return the computed value or raise an <a class="reference internal" href="../library/exceptions#AttributeError" title="AttributeError"><code>AttributeError</code></a>. If an attribute is not found on a module object through the normal lookup, i.e. <a class="reference internal" href="#object.__getattribute__" title="object.__getattribute__"><code>object.__getattribute__()</code></a>, then <code>__getattr__</code> is searched in the module <code>__dict__</code> before raising an <a class="reference internal" href="../library/exceptions#AttributeError" title="AttributeError"><code>AttributeError</code></a>. If found, it is called with the attribute name and the result is returned.</p> <p>The <code>__dir__</code> function should accept no arguments, and return a sequence of strings that represents the names accessible on module. If present, this function overrides the standard <a class="reference internal" href="../library/functions#dir" title="dir"><code>dir()</code></a> search on a module.</p> <p>For a more fine grained customization of the module behavior (setting attributes, properties, etc.), one can set the <code>__class__</code> attribute of a module object to a subclass of <a class="reference internal" href="../library/types#types.ModuleType" title="types.ModuleType"><code>types.ModuleType</code></a>. For example:</p> <pre data-language="python">import sys +from types import ModuleType + +class VerboseModule(ModuleType): + def __repr__(self): + return f'Verbose {self.__name__}' + + def __setattr__(self, attr, value): + print(f'Setting {attr}...') + super().__setattr__(attr, value) + +sys.modules[__name__].__class__ = VerboseModule +</pre> <div class="admonition note"> <p class="admonition-title">Note</p> <p>Defining module <code>__getattr__</code> and setting module <code>__class__</code> only affect lookups made using the attribute access syntax – directly accessing the module globals (whether by code within the module, or via a reference to the module’s globals dictionary) is unaffected.</p> </div> <div class="versionchanged"> <p><span class="versionmodified changed">Changed in version 3.5: </span><code>__class__</code> module attribute is now writable.</p> </div> <div class="versionadded"> <p><span class="versionmodified added">New in version 3.7: </span><code>__getattr__</code> and <code>__dir__</code> module attributes.</p> </div> <div class="admonition seealso"> <p class="admonition-title">See also</p> <dl class="simple"> <dt> +<span class="target" id="index-84"></span><a class="pep reference external" href="https://peps.python.org/pep-0562/"><strong>PEP 562</strong></a> - Module __getattr__ and __dir__</dt> +<dd> +<p>Describes the <code>__getattr__</code> and <code>__dir__</code> functions on modules.</p> </dd> </dl> </div> </section> <section id="implementing-descriptors"> <span id="descriptors"></span><h4> +<span class="section-number">3.3.2.2. </span>Implementing Descriptors</h4> <p>The following methods only apply when an instance of the class containing the method (a so-called <em>descriptor</em> class) appears in an <em>owner</em> class (the descriptor must be in either the owner’s class dictionary or in the class dictionary for one of its parents). In the examples below, “the attribute” refers to the attribute whose name is the key of the property in the owner class’ <a class="reference internal" href="../library/stdtypes#object.__dict__" title="object.__dict__"><code>__dict__</code></a>.</p> <dl class="py method"> <dt class="sig sig-object py" id="object.__get__"> +<code>object.__get__(self, instance, owner=None)</code> </dt> <dd> +<p>Called to get the attribute of the owner class (class attribute access) or of an instance of that class (instance attribute access). The optional <em>owner</em> argument is the owner class, while <em>instance</em> is the instance that the attribute was accessed through, or <code>None</code> when the attribute is accessed through the <em>owner</em>.</p> <p>This method should return the computed attribute value or raise an <a class="reference internal" href="../library/exceptions#AttributeError" title="AttributeError"><code>AttributeError</code></a> exception.</p> <p><span class="target" id="index-85"></span><a class="pep reference external" href="https://peps.python.org/pep-0252/"><strong>PEP 252</strong></a> specifies that <a class="reference internal" href="#object.__get__" title="object.__get__"><code>__get__()</code></a> is callable with one or two arguments. Python’s own built-in descriptors support this specification; however, it is likely that some third-party tools have descriptors that require both arguments. Python’s own <a class="reference internal" href="#object.__getattribute__" title="object.__getattribute__"><code>__getattribute__()</code></a> implementation always passes in both arguments whether they are required or not.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__set__"> +<code>object.__set__(self, instance, value)</code> </dt> <dd> +<p>Called to set the attribute on an instance <em>instance</em> of the owner class to a new value, <em>value</em>.</p> <p>Note, adding <a class="reference internal" href="#object.__set__" title="object.__set__"><code>__set__()</code></a> or <a class="reference internal" href="#object.__delete__" title="object.__delete__"><code>__delete__()</code></a> changes the kind of descriptor to a “data descriptor”. See <a class="reference internal" href="#descriptor-invocation"><span class="std std-ref">Invoking Descriptors</span></a> for more details.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__delete__"> +<code>object.__delete__(self, instance)</code> </dt> <dd> +<p>Called to delete the attribute on an instance <em>instance</em> of the owner class.</p> </dd> +</dl> <p>Instances of descriptors may also have the <code>__objclass__</code> attribute present:</p> <dl class="py attribute"> <dt class="sig sig-object py" id="object.__objclass__"> +<code>object.__objclass__</code> </dt> <dd> +<p>The attribute <code>__objclass__</code> is interpreted by the <a class="reference internal" href="../library/inspect#module-inspect" title="inspect: Extract information and source code from live objects."><code>inspect</code></a> module as specifying the class where this object was defined (setting this appropriately can assist in runtime introspection of dynamic class attributes). For callables, it may indicate that an instance of the given type (or a subclass) is expected or required as the first positional argument (for example, CPython sets this attribute for unbound methods that are implemented in C).</p> </dd> +</dl> </section> <section id="invoking-descriptors"> <span id="descriptor-invocation"></span><h4> +<span class="section-number">3.3.2.3. </span>Invoking Descriptors</h4> <p>In general, a descriptor is an object attribute with “binding behavior”, one whose attribute access has been overridden by methods in the descriptor protocol: <a class="reference internal" href="#object.__get__" title="object.__get__"><code>__get__()</code></a>, <a class="reference internal" href="#object.__set__" title="object.__set__"><code>__set__()</code></a>, and <a class="reference internal" href="#object.__delete__" title="object.__delete__"><code>__delete__()</code></a>. If any of those methods are defined for an object, it is said to be a descriptor.</p> <p>The default behavior for attribute access is to get, set, or delete the attribute from an object’s dictionary. For instance, <code>a.x</code> has a lookup chain starting with <code>a.__dict__['x']</code>, then <code>type(a).__dict__['x']</code>, and continuing through the base classes of <code>type(a)</code> excluding metaclasses.</p> <p>However, if the looked-up value is an object defining one of the descriptor methods, then Python may override the default behavior and invoke the descriptor method instead. Where this occurs in the precedence chain depends on which descriptor methods were defined and how they were called.</p> <p>The starting point for descriptor invocation is a binding, <code>a.x</code>. How the arguments are assembled depends on <code>a</code>:</p> <dl class="simple"> <dt>Direct Call</dt> +<dd> +<p>The simplest and least common call is when user code directly invokes a descriptor method: <code>x.__get__(a)</code>.</p> </dd> <dt>Instance Binding</dt> +<dd> +<p>If binding to an object instance, <code>a.x</code> is transformed into the call: <code>type(a).__dict__['x'].__get__(a, type(a))</code>.</p> </dd> <dt>Class Binding</dt> +<dd> +<p>If binding to a class, <code>A.x</code> is transformed into the call: <code>A.__dict__['x'].__get__(None, A)</code>.</p> </dd> <dt>Super Binding</dt> +<dd> +<p>A dotted lookup such as <code>super(A, a).x</code> searches <code>a.__class__.__mro__</code> for a base class <code>B</code> following <code>A</code> and then returns <code>B.__dict__['x'].__get__(a, A)</code>. If not a descriptor, <code>x</code> is returned unchanged.</p> </dd> </dl> <p>For instance bindings, the precedence of descriptor invocation depends on which descriptor methods are defined. A descriptor can define any combination of <a class="reference internal" href="#object.__get__" title="object.__get__"><code>__get__()</code></a>, <a class="reference internal" href="#object.__set__" title="object.__set__"><code>__set__()</code></a> and <a class="reference internal" href="#object.__delete__" title="object.__delete__"><code>__delete__()</code></a>. If it does not define <code>__get__()</code>, then accessing the attribute will return the descriptor object itself unless there is a value in the object’s instance dictionary. If the descriptor defines <code>__set__()</code> and/or <code>__delete__()</code>, it is a data descriptor; if it defines neither, it is a non-data descriptor. Normally, data descriptors define both <code>__get__()</code> and <code>__set__()</code>, while non-data descriptors have just the <code>__get__()</code> method. Data descriptors with <code>__get__()</code> and <code>__set__()</code> (and/or <code>__delete__()</code>) defined always override a redefinition in an instance dictionary. In contrast, non-data descriptors can be overridden by instances.</p> <p>Python methods (including those decorated with <a class="reference internal" href="../library/functions#staticmethod" title="staticmethod"><code>@staticmethod</code></a> and <a class="reference internal" href="../library/functions#classmethod" title="classmethod"><code>@classmethod</code></a>) are implemented as non-data descriptors. Accordingly, instances can redefine and override methods. This allows individual instances to acquire behaviors that differ from other instances of the same class.</p> <p>The <a class="reference internal" href="../library/functions#property" title="property"><code>property()</code></a> function is implemented as a data descriptor. Accordingly, instances cannot override the behavior of a property.</p> </section> <section id="slots"> <span id="id11"></span><h4> +<span class="section-number">3.3.2.4. </span>__slots__</h4> <p><em>__slots__</em> allow us to explicitly declare data members (like properties) and deny the creation of <a class="reference internal" href="../library/stdtypes#object.__dict__" title="object.__dict__"><code>__dict__</code></a> and <em>__weakref__</em> (unless explicitly declared in <em>__slots__</em> or available in a parent.)</p> <p>The space saved over using <a class="reference internal" href="../library/stdtypes#object.__dict__" title="object.__dict__"><code>__dict__</code></a> can be significant. Attribute lookup speed can be significantly improved as well.</p> <dl class="py data"> <dt class="sig sig-object py" id="object.__slots__"> +<code>object.__slots__</code> </dt> <dd> +<p>This class variable can be assigned a string, iterable, or sequence of strings with variable names used by instances. <em>__slots__</em> reserves space for the declared variables and prevents the automatic creation of <a class="reference internal" href="../library/stdtypes#object.__dict__" title="object.__dict__"><code>__dict__</code></a> and <em>__weakref__</em> for each instance.</p> </dd> +</dl> <p id="datamodel-note-slots">Notes on using <em>__slots__</em>:</p> <ul class="simple"> <li>When inheriting from a class without <em>__slots__</em>, the <a class="reference internal" href="../library/stdtypes#object.__dict__" title="object.__dict__"><code>__dict__</code></a> and <em>__weakref__</em> attribute of the instances will always be accessible.</li> <li>Without a <a class="reference internal" href="../library/stdtypes#object.__dict__" title="object.__dict__"><code>__dict__</code></a> variable, instances cannot be assigned new variables not listed in the <em>__slots__</em> definition. Attempts to assign to an unlisted variable name raises <a class="reference internal" href="../library/exceptions#AttributeError" title="AttributeError"><code>AttributeError</code></a>. If dynamic assignment of new variables is desired, then add <code>'__dict__'</code> to the sequence of strings in the <em>__slots__</em> declaration.</li> <li>Without a <em>__weakref__</em> variable for each instance, classes defining <em>__slots__</em> do not support <a class="reference internal" href="../library/weakref#module-weakref" title="weakref: Support for weak references and weak dictionaries."><code>weak references</code></a> to its instances. If weak reference support is needed, then add <code>'__weakref__'</code> to the sequence of strings in the <em>__slots__</em> declaration.</li> <li> +<em>__slots__</em> are implemented at the class level by creating <a class="reference internal" href="#descriptors"><span class="std std-ref">descriptors</span></a> for each variable name. As a result, class attributes cannot be used to set default values for instance variables defined by <em>__slots__</em>; otherwise, the class attribute would overwrite the descriptor assignment.</li> <li>The action of a <em>__slots__</em> declaration is not limited to the class where it is defined. <em>__slots__</em> declared in parents are available in child classes. However, child subclasses will get a <a class="reference internal" href="../library/stdtypes#object.__dict__" title="object.__dict__"><code>__dict__</code></a> and <em>__weakref__</em> unless they also define <em>__slots__</em> (which should only contain names of any <em>additional</em> slots).</li> <li>If a class defines a slot also defined in a base class, the instance variable defined by the base class slot is inaccessible (except by retrieving its descriptor directly from the base class). This renders the meaning of the program undefined. In the future, a check may be added to prevent this.</li> <li> +<a class="reference internal" href="../library/exceptions#TypeError" title="TypeError"><code>TypeError</code></a> will be raised if nonempty <em>__slots__</em> are defined for a class derived from a <a class="reference internal" href="../c-api/typeobj#c.PyTypeObject.tp_itemsize" title="PyTypeObject.tp_itemsize"><code>"variable-length" built-in type</code></a> such as <a class="reference internal" href="../library/functions#int" title="int"><code>int</code></a>, <a class="reference internal" href="../library/stdtypes#bytes" title="bytes"><code>bytes</code></a>, and <a class="reference internal" href="../library/stdtypes#tuple" title="tuple"><code>tuple</code></a>.</li> <li>Any non-string <a class="reference internal" href="../glossary#term-iterable"><span class="xref std std-term">iterable</span></a> may be assigned to <em>__slots__</em>.</li> <li>If a <a class="reference internal" href="../library/stdtypes#dict" title="dict"><code>dictionary</code></a> is used to assign <em>__slots__</em>, the dictionary keys will be used as the slot names. The values of the dictionary can be used to provide per-attribute docstrings that will be recognised by <a class="reference internal" href="../library/inspect#inspect.getdoc" title="inspect.getdoc"><code>inspect.getdoc()</code></a> and displayed in the output of <a class="reference internal" href="../library/functions#help" title="help"><code>help()</code></a>.</li> <li> +<a class="reference internal" href="../library/stdtypes#instance.__class__" title="instance.__class__"><code>__class__</code></a> assignment works only if both classes have the same <em>__slots__</em>.</li> <li> +<a class="reference internal" href="../tutorial/classes#tut-multiple"><span class="std std-ref">Multiple inheritance</span></a> with multiple slotted parent classes can be used, but only one parent is allowed to have attributes created by slots (the other bases must have empty slot layouts) - violations raise <a class="reference internal" href="../library/exceptions#TypeError" title="TypeError"><code>TypeError</code></a>.</li> <li>If an <a class="reference internal" href="../glossary#term-iterator"><span class="xref std std-term">iterator</span></a> is used for <em>__slots__</em> then a <a class="reference internal" href="../glossary#term-descriptor"><span class="xref std std-term">descriptor</span></a> is created for each of the iterator’s values. However, the <em>__slots__</em> attribute will be an empty iterator.</li> </ul> </section> </section> <section id="customizing-class-creation"> <span id="class-customization"></span><h3> +<span class="section-number">3.3.3. </span>Customizing class creation</h3> <p>Whenever a class inherits from another class, <a class="reference internal" href="#object.__init_subclass__" title="object.__init_subclass__"><code>__init_subclass__()</code></a> is called on the parent class. This way, it is possible to write classes which change the behavior of subclasses. This is closely related to class decorators, but where class decorators only affect the specific class they’re applied to, <code>__init_subclass__</code> solely applies to future subclasses of the class defining the method.</p> <dl class="py method"> <dt class="sig sig-object py" id="object.__init_subclass__"> +<code>classmethod object.__init_subclass__(cls)</code> </dt> <dd> +<p>This method is called whenever the containing class is subclassed. <em>cls</em> is then the new subclass. If defined as a normal instance method, this method is implicitly converted to a class method.</p> <p>Keyword arguments which are given to a new class are passed to the parent’s class <code>__init_subclass__</code>. For compatibility with other classes using <code>__init_subclass__</code>, one should take out the needed keyword arguments and pass the others over to the base class, as in:</p> <pre data-language="python">class Philosopher: + def __init_subclass__(cls, /, default_name, **kwargs): + super().__init_subclass__(**kwargs) + cls.default_name = default_name + +class AustralianPhilosopher(Philosopher, default_name="Bruce"): + pass +</pre> <p>The default implementation <code>object.__init_subclass__</code> does nothing, but raises an error if it is called with any arguments.</p> <div class="admonition note"> <p class="admonition-title">Note</p> <p>The metaclass hint <code>metaclass</code> is consumed by the rest of the type machinery, and is never passed to <code>__init_subclass__</code> implementations. The actual metaclass (rather than the explicit hint) can be accessed as <code>type(cls)</code>.</p> </div> <div class="versionadded"> <p><span class="versionmodified added">New in version 3.6.</span></p> </div> </dd> +</dl> <p>When a class is created, <code>type.__new__()</code> scans the class variables and makes callbacks to those with a <a class="reference internal" href="#object.__set_name__" title="object.__set_name__"><code>__set_name__()</code></a> hook.</p> <dl class="py method"> <dt class="sig sig-object py" id="object.__set_name__"> +<code>object.__set_name__(self, owner, name)</code> </dt> <dd> +<p>Automatically called at the time the owning class <em>owner</em> is created. The object has been assigned to <em>name</em> in that class:</p> <pre data-language="python">class A: + x = C() # Automatically calls: x.__set_name__(A, 'x') +</pre> <p>If the class variable is assigned after the class is created, <a class="reference internal" href="#object.__set_name__" title="object.__set_name__"><code>__set_name__()</code></a> will not be called automatically. If needed, <a class="reference internal" href="#object.__set_name__" title="object.__set_name__"><code>__set_name__()</code></a> can be called directly:</p> <pre data-language="python">class A: + pass + +c = C() +A.x = c # The hook is not called +c.__set_name__(A, 'x') # Manually invoke the hook +</pre> <p>See <a class="reference internal" href="#class-object-creation"><span class="std std-ref">Creating the class object</span></a> for more details.</p> <div class="versionadded"> <p><span class="versionmodified added">New in version 3.6.</span></p> </div> </dd> +</dl> <section id="metaclasses"> <span id="id12"></span><h4> +<span class="section-number">3.3.3.1. </span>Metaclasses</h4> <p id="index-86">By default, classes are constructed using <a class="reference internal" href="../library/functions#type" title="type"><code>type()</code></a>. The class body is executed in a new namespace and the class name is bound locally to the result of <code>type(name, bases, namespace)</code>.</p> <p>The class creation process can be customized by passing the <code>metaclass</code> keyword argument in the class definition line, or by inheriting from an existing class that included such an argument. In the following example, both <code>MyClass</code> and <code>MySubclass</code> are instances of <code>Meta</code>:</p> <pre data-language="python">class Meta(type): + pass + +class MyClass(metaclass=Meta): + pass + +class MySubclass(MyClass): + pass +</pre> <p>Any other keyword arguments that are specified in the class definition are passed through to all metaclass operations described below.</p> <p>When a class definition is executed, the following steps occur:</p> <ul class="simple"> <li>MRO entries are resolved;</li> <li>the appropriate metaclass is determined;</li> <li>the class namespace is prepared;</li> <li>the class body is executed;</li> <li>the class object is created.</li> </ul> </section> <section id="resolving-mro-entries"> <h4> +<span class="section-number">3.3.3.2. </span>Resolving MRO entries</h4> <dl class="py method"> <dt class="sig sig-object py" id="object.__mro_entries__"> +<code>object.__mro_entries__(self, bases)</code> </dt> <dd> +<p>If a base that appears in a class definition is not an instance of <a class="reference internal" href="../library/functions#type" title="type"><code>type</code></a>, then an <code>__mro_entries__()</code> method is searched on the base. If an <code>__mro_entries__()</code> method is found, the base is substituted with the result of a call to <code>__mro_entries__()</code> when creating the class. The method is called with the original bases tuple passed to the <em>bases</em> parameter, and must return a tuple of classes that will be used instead of the base. The returned tuple may be empty: in these cases, the original base is ignored.</p> </dd> +</dl> <div class="admonition seealso"> <p class="admonition-title">See also</p> <dl class="simple"> <dt> +<code></code> <a class="reference internal" href="../library/types#types.resolve_bases" title="types.resolve_bases"><code>types.resolve_bases()</code></a> +</dt> +<dd> +<p>Dynamically resolve bases that are not instances of <a class="reference internal" href="../library/functions#type" title="type"><code>type</code></a>.</p> </dd> <dt> +<code></code> <a class="reference internal" href="../library/types#types.get_original_bases" title="types.get_original_bases"><code>types.get_original_bases()</code></a> +</dt> +<dd> +<p>Retrieve a class’s “original bases” prior to modifications by <a class="reference internal" href="#object.__mro_entries__" title="object.__mro_entries__"><code>__mro_entries__()</code></a>.</p> </dd> <dt> +<span class="target" id="index-87"></span><a class="pep reference external" href="https://peps.python.org/pep-0560/"><strong>PEP 560</strong></a> +</dt> +<dd> +<p>Core support for typing module and generic types.</p> </dd> </dl> </div> </section> <section id="determining-the-appropriate-metaclass"> <h4> +<span class="section-number">3.3.3.3. </span>Determining the appropriate metaclass</h4> <p id="index-88">The appropriate metaclass for a class definition is determined as follows:</p> <ul class="simple"> <li>if no bases and no explicit metaclass are given, then <a class="reference internal" href="../library/functions#type" title="type"><code>type()</code></a> is used;</li> <li>if an explicit metaclass is given and it is <em>not</em> an instance of <a class="reference internal" href="../library/functions#type" title="type"><code>type()</code></a>, then it is used directly as the metaclass;</li> <li>if an instance of <a class="reference internal" href="../library/functions#type" title="type"><code>type()</code></a> is given as the explicit metaclass, or bases are defined, then the most derived metaclass is used.</li> </ul> <p>The most derived metaclass is selected from the explicitly specified metaclass (if any) and the metaclasses (i.e. <code>type(cls)</code>) of all specified base classes. The most derived metaclass is one which is a subtype of <em>all</em> of these candidate metaclasses. If none of the candidate metaclasses meets that criterion, then the class definition will fail with <code>TypeError</code>.</p> </section> <section id="preparing-the-class-namespace"> <span id="prepare"></span><h4> +<span class="section-number">3.3.3.4. </span>Preparing the class namespace</h4> <p id="index-89">Once the appropriate metaclass has been identified, then the class namespace is prepared. If the metaclass has a <code>__prepare__</code> attribute, it is called as <code>namespace = metaclass.__prepare__(name, bases, **kwds)</code> (where the additional keyword arguments, if any, come from the class definition). The <code>__prepare__</code> method should be implemented as a <a class="reference internal" href="../library/functions#classmethod" title="classmethod"><code>classmethod</code></a>. The namespace returned by <code>__prepare__</code> is passed in to <code>__new__</code>, but when the final class object is created the namespace is copied into a new <code>dict</code>.</p> <p>If the metaclass has no <code>__prepare__</code> attribute, then the class namespace is initialised as an empty ordered mapping.</p> <div class="admonition seealso"> <p class="admonition-title">See also</p> <dl class="simple"> <dt> +<span class="target" id="index-90"></span><a class="pep reference external" href="https://peps.python.org/pep-3115/"><strong>PEP 3115</strong></a> - Metaclasses in Python 3000</dt> +<dd> +<p>Introduced the <code>__prepare__</code> namespace hook</p> </dd> </dl> </div> </section> <section id="executing-the-class-body"> <h4> +<span class="section-number">3.3.3.5. </span>Executing the class body</h4> <p id="index-91">The class body is executed (approximately) as <code>exec(body, globals(), namespace)</code>. The key difference from a normal call to <a class="reference internal" href="../library/functions#exec" title="exec"><code>exec()</code></a> is that lexical scoping allows the class body (including any methods) to reference names from the current and outer scopes when the class definition occurs inside a function.</p> <p>However, even when the class definition occurs inside the function, methods defined inside the class still cannot see names defined at the class scope. Class variables must be accessed through the first parameter of instance or class methods, or through the implicit lexically scoped <code>__class__</code> reference described in the next section.</p> </section> <section id="creating-the-class-object"> <span id="class-object-creation"></span><h4> +<span class="section-number">3.3.3.6. </span>Creating the class object</h4> <p id="index-92">Once the class namespace has been populated by executing the class body, the class object is created by calling <code>metaclass(name, bases, namespace, **kwds)</code> (the additional keywords passed here are the same as those passed to <code>__prepare__</code>).</p> <p>This class object is the one that will be referenced by the zero-argument form of <a class="reference internal" href="../library/functions#super" title="super"><code>super()</code></a>. <code>__class__</code> is an implicit closure reference created by the compiler if any methods in a class body refer to either <code>__class__</code> or <code>super</code>. This allows the zero argument form of <a class="reference internal" href="../library/functions#super" title="super"><code>super()</code></a> to correctly identify the class being defined based on lexical scoping, while the class or instance that was used to make the current call is identified based on the first argument passed to the method.</p> <div class="impl-detail compound"> <p><strong>CPython implementation detail:</strong> In CPython 3.6 and later, the <code>__class__</code> cell is passed to the metaclass as a <code>__classcell__</code> entry in the class namespace. If present, this must be propagated up to the <code>type.__new__</code> call in order for the class to be initialised correctly. Failing to do so will result in a <a class="reference internal" href="../library/exceptions#RuntimeError" title="RuntimeError"><code>RuntimeError</code></a> in Python 3.8.</p> </div> <p>When using the default metaclass <a class="reference internal" href="../library/functions#type" title="type"><code>type</code></a>, or any metaclass that ultimately calls <code>type.__new__</code>, the following additional customization steps are invoked after creating the class object:</p> <ol class="arabic simple"> <li>The <code>type.__new__</code> method collects all of the attributes in the class namespace that define a <a class="reference internal" href="#object.__set_name__" title="object.__set_name__"><code>__set_name__()</code></a> method;</li> <li>Those <code>__set_name__</code> methods are called with the class being defined and the assigned name of that particular attribute;</li> <li>The <a class="reference internal" href="#object.__init_subclass__" title="object.__init_subclass__"><code>__init_subclass__()</code></a> hook is called on the immediate parent of the new class in its method resolution order.</li> </ol> <p>After the class object is created, it is passed to the class decorators included in the class definition (if any) and the resulting object is bound in the local namespace as the defined class.</p> <p>When a new class is created by <code>type.__new__</code>, the object provided as the namespace parameter is copied to a new ordered mapping and the original object is discarded. The new copy is wrapped in a read-only proxy, which becomes the <a class="reference internal" href="../library/stdtypes#object.__dict__" title="object.__dict__"><code>__dict__</code></a> attribute of the class object.</p> <div class="admonition seealso"> <p class="admonition-title">See also</p> <dl class="simple"> <dt> +<span class="target" id="index-93"></span><a class="pep reference external" href="https://peps.python.org/pep-3135/"><strong>PEP 3135</strong></a> - New super</dt> +<dd> +<p>Describes the implicit <code>__class__</code> closure reference</p> </dd> </dl> </div> </section> <section id="uses-for-metaclasses"> <h4> +<span class="section-number">3.3.3.7. </span>Uses for metaclasses</h4> <p>The potential uses for metaclasses are boundless. Some ideas that have been explored include enum, logging, interface checking, automatic delegation, automatic property creation, proxies, frameworks, and automatic resource locking/synchronization.</p> </section> </section> <section id="customizing-instance-and-subclass-checks"> <h3> +<span class="section-number">3.3.4. </span>Customizing instance and subclass checks</h3> <p>The following methods are used to override the default behavior of the <a class="reference internal" href="../library/functions#isinstance" title="isinstance"><code>isinstance()</code></a> and <a class="reference internal" href="../library/functions#issubclass" title="issubclass"><code>issubclass()</code></a> built-in functions.</p> <p>In particular, the metaclass <a class="reference internal" href="../library/abc#abc.ABCMeta" title="abc.ABCMeta"><code>abc.ABCMeta</code></a> implements these methods in order to allow the addition of Abstract Base Classes (ABCs) as “virtual base classes” to any class or type (including built-in types), including other ABCs.</p> <dl class="py method"> <dt class="sig sig-object py" id="class.__instancecheck__"> +<code>class.__instancecheck__(self, instance)</code> </dt> <dd> +<p>Return true if <em>instance</em> should be considered a (direct or indirect) instance of <em>class</em>. If defined, called to implement <code>isinstance(instance, +class)</code>.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="class.__subclasscheck__"> +<code>class.__subclasscheck__(self, subclass)</code> </dt> <dd> +<p>Return true if <em>subclass</em> should be considered a (direct or indirect) subclass of <em>class</em>. If defined, called to implement <code>issubclass(subclass, +class)</code>.</p> </dd> +</dl> <p>Note that these methods are looked up on the type (metaclass) of a class. They cannot be defined as class methods in the actual class. This is consistent with the lookup of special methods that are called on instances, only in this case the instance is itself a class.</p> <div class="admonition seealso"> <p class="admonition-title">See also</p> <dl class="simple"> <dt> +<span class="target" id="index-94"></span><a class="pep reference external" href="https://peps.python.org/pep-3119/"><strong>PEP 3119</strong></a> - Introducing Abstract Base Classes</dt> +<dd> +<p>Includes the specification for customizing <a class="reference internal" href="../library/functions#isinstance" title="isinstance"><code>isinstance()</code></a> and <a class="reference internal" href="../library/functions#issubclass" title="issubclass"><code>issubclass()</code></a> behavior through <a class="reference internal" href="#class.__instancecheck__" title="class.__instancecheck__"><code>__instancecheck__()</code></a> and <a class="reference internal" href="#class.__subclasscheck__" title="class.__subclasscheck__"><code>__subclasscheck__()</code></a>, with motivation for this functionality in the context of adding Abstract Base Classes (see the <a class="reference internal" href="../library/abc#module-abc" title="abc: Abstract base classes according to :pep:`3119`."><code>abc</code></a> module) to the language.</p> </dd> </dl> </div> </section> <section id="emulating-generic-types"> <h3> +<span class="section-number">3.3.5. </span>Emulating generic types</h3> <p>When using <a class="reference internal" href="../glossary#term-annotation"><span class="xref std std-term">type annotations</span></a>, it is often useful to <em>parameterize</em> a <a class="reference internal" href="../glossary#term-generic-type"><span class="xref std std-term">generic type</span></a> using Python’s square-brackets notation. For example, the annotation <code>list[int]</code> might be used to signify a <a class="reference internal" href="../library/stdtypes#list" title="list"><code>list</code></a> in which all the elements are of type <a class="reference internal" href="../library/functions#int" title="int"><code>int</code></a>.</p> <div class="admonition seealso"> <p class="admonition-title">See also</p> <dl class="simple"> <dt> +<span class="target" id="index-95"></span><a class="pep reference external" href="https://peps.python.org/pep-0484/"><strong>PEP 484</strong></a> - Type Hints</dt> +<dd> +<p>Introducing Python’s framework for type annotations</p> </dd> <dt><a class="reference internal" href="../library/stdtypes#types-genericalias"><span class="std std-ref">Generic Alias Types</span></a></dt> +<dd> +<p>Documentation for objects representing parameterized generic classes</p> </dd> <dt> +<code>Generics, user-defined generics and</code> <a class="reference internal" href="../library/typing#typing.Generic" title="typing.Generic"><code>typing.Generic</code></a> +</dt> +<dd> +<p>Documentation on how to implement generic classes that can be parameterized at runtime and understood by static type-checkers.</p> </dd> </dl> </div> <p>A class can <em>generally</em> only be parameterized if it defines the special class method <code>__class_getitem__()</code>.</p> <dl class="py method"> <dt class="sig sig-object py" id="object.__class_getitem__"> +<code>classmethod object.__class_getitem__(cls, key)</code> </dt> <dd> +<p>Return an object representing the specialization of a generic class by type arguments found in <em>key</em>.</p> <p>When defined on a class, <code>__class_getitem__()</code> is automatically a class method. As such, there is no need for it to be decorated with <a class="reference internal" href="../library/functions#classmethod" title="classmethod"><code>@classmethod</code></a> when it is defined.</p> </dd> +</dl> <section id="the-purpose-of-class-getitem"> <h4> +<span class="section-number">3.3.5.1. </span>The purpose of <em>__class_getitem__</em> +</h4> <p>The purpose of <a class="reference internal" href="#object.__class_getitem__" title="object.__class_getitem__"><code>__class_getitem__()</code></a> is to allow runtime parameterization of standard-library generic classes in order to more easily apply <a class="reference internal" href="../glossary#term-type-hint"><span class="xref std std-term">type hints</span></a> to these classes.</p> <p>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 <a class="reference internal" href="#object.__class_getitem__" title="object.__class_getitem__"><code>__class_getitem__()</code></a>, or inherit from <a class="reference internal" href="../library/typing#typing.Generic" title="typing.Generic"><code>typing.Generic</code></a>, which has its own implementation of <code>__class_getitem__()</code>.</p> <p>Custom implementations of <a class="reference internal" href="#object.__class_getitem__" title="object.__class_getitem__"><code>__class_getitem__()</code></a> on classes defined outside of the standard library may not be understood by third-party type-checkers such as mypy. Using <code>__class_getitem__()</code> on any class for purposes other than type hinting is discouraged.</p> </section> <section id="class-getitem-versus-getitem"> <span id="classgetitem-versus-getitem"></span><h4> +<span class="section-number">3.3.5.2. </span><em>__class_getitem__</em> versus <em>__getitem__</em> +</h4> <p>Usually, the <a class="reference internal" href="expressions#subscriptions"><span class="std std-ref">subscription</span></a> of an object using square brackets will call the <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><code>__getitem__()</code></a> instance method defined on the object’s class. However, if the object being subscribed is itself a class, the class method <a class="reference internal" href="#object.__class_getitem__" title="object.__class_getitem__"><code>__class_getitem__()</code></a> may be called instead. <code>__class_getitem__()</code> should return a <a class="reference internal" href="../library/stdtypes#types-genericalias"><span class="std std-ref">GenericAlias</span></a> object if it is properly defined.</p> <p>Presented with the <a class="reference internal" href="../glossary#term-expression"><span class="xref std std-term">expression</span></a> <code>obj[x]</code>, the Python interpreter follows something like the following process to decide whether <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><code>__getitem__()</code></a> or <a class="reference internal" href="#object.__class_getitem__" title="object.__class_getitem__"><code>__class_getitem__()</code></a> should be called:</p> <pre data-language="python">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" + ) +</pre> <p>In Python, all classes are themselves instances of other classes. The class of a class is known as that class’s <a class="reference internal" href="../glossary#term-metaclass"><span class="xref std std-term">metaclass</span></a>, and most classes have the <a class="reference internal" href="../library/functions#type" title="type"><code>type</code></a> class as their metaclass. <a class="reference internal" href="../library/functions#type" title="type"><code>type</code></a> does not define <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><code>__getitem__()</code></a>, meaning that expressions such as <code>list[int]</code>, <code>dict[str, float]</code> and <code>tuple[str, bytes]</code> all result in <a class="reference internal" href="#object.__class_getitem__" title="object.__class_getitem__"><code>__class_getitem__()</code></a> being called:</p> <pre data-language="python">>>> # 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'> +</pre> <p>However, if a class has a custom metaclass that defines <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><code>__getitem__()</code></a>, subscribing the class may result in different behaviour. An example of this can be found in the <a class="reference internal" href="../library/enum#module-enum" title="enum: Implementation of an enumeration class."><code>enum</code></a> module:</p> <pre data-language="python">>>> 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'> +</pre> <div class="admonition seealso"> <p class="admonition-title">See also</p> <dl class="simple"> <dt> +<span class="target" id="index-96"></span><a class="pep reference external" href="https://peps.python.org/pep-0560/"><strong>PEP 560</strong></a> - Core Support for typing module and generic types</dt> +<dd> +<p>Introducing <a class="reference internal" href="#object.__class_getitem__" title="object.__class_getitem__"><code>__class_getitem__()</code></a>, and outlining when a <a class="reference internal" href="expressions#subscriptions"><span class="std std-ref">subscription</span></a> results in <code>__class_getitem__()</code> being called instead of <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><code>__getitem__()</code></a></p> </dd> </dl> </div> </section> </section> <section id="emulating-callable-objects"> <span id="id13"></span><h3> +<span class="section-number">3.3.6. </span>Emulating callable objects</h3> <dl class="py method"> <dt class="sig sig-object py" id="object.__call__"> +<code>object.__call__(self[, args...])</code> </dt> <dd> +<p id="index-97">Called when the instance is “called” as a function; if this method is defined, <code>x(arg1, arg2, ...)</code> roughly translates to <code>type(x).__call__(x, arg1, ...)</code>.</p> </dd> +</dl> </section> <section id="emulating-container-types"> <span id="sequence-types"></span><h3> +<span class="section-number">3.3.7. </span>Emulating container types</h3> <p>The following methods can be defined to implement container objects. Containers usually are <a class="reference internal" href="../glossary#term-sequence"><span class="xref std std-term">sequences</span></a> (such as <a class="reference internal" href="../library/stdtypes#list" title="list"><code>lists</code></a> or <a class="reference internal" href="../library/stdtypes#tuple" title="tuple"><code>tuples</code></a>) or <a class="reference internal" href="../glossary#term-mapping"><span class="xref std std-term">mappings</span></a> (like <a class="reference internal" href="../library/stdtypes#dict" title="dict"><code>dictionaries</code></a>), but can represent other containers as well. The first set of methods is used either to emulate a sequence or to emulate a mapping; the difference is that for a sequence, the allowable keys should be the integers <em>k</em> for which <code>0 <= k < +N</code> where <em>N</em> is the length of the sequence, or <a class="reference internal" href="../library/functions#slice" title="slice"><code>slice</code></a> objects, which define a range of items. It is also recommended that mappings provide the methods <code>keys()</code>, <code>values()</code>, <code>items()</code>, <code>get()</code>, <code>clear()</code>, <code>setdefault()</code>, <code>pop()</code>, <code>popitem()</code>, <code>copy()</code>, and <code>update()</code> behaving similar to those for Python’s standard <a class="reference internal" href="../library/stdtypes#dict" title="dict"><code>dictionary</code></a> objects. The <a class="reference internal" href="../library/collections.abc#module-collections.abc" title="collections.abc: Abstract base classes for containers"><code>collections.abc</code></a> module provides a <a class="reference internal" href="../library/collections.abc#collections.abc.MutableMapping" title="collections.abc.MutableMapping"><code>MutableMapping</code></a> <a class="reference internal" href="../glossary#term-abstract-base-class"><span class="xref std std-term">abstract base class</span></a> to help create those methods from a base set of <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><code>__getitem__()</code></a>, <a class="reference internal" href="#object.__setitem__" title="object.__setitem__"><code>__setitem__()</code></a>, <a class="reference internal" href="#object.__delitem__" title="object.__delitem__"><code>__delitem__()</code></a>, and <code>keys()</code>. Mutable sequences should provide methods <code>append()</code>, <code>count()</code>, <code>index()</code>, <code>extend()</code>, <code>insert()</code>, <code>pop()</code>, <code>remove()</code>, <code>reverse()</code> and <code>sort()</code>, like Python standard <a class="reference internal" href="../library/stdtypes#list" title="list"><code>list</code></a> objects. Finally, sequence types should implement addition (meaning concatenation) and multiplication (meaning repetition) by defining the methods <a class="reference internal" href="#object.__add__" title="object.__add__"><code>__add__()</code></a>, <a class="reference internal" href="#object.__radd__" title="object.__radd__"><code>__radd__()</code></a>, <a class="reference internal" href="#object.__iadd__" title="object.__iadd__"><code>__iadd__()</code></a>, <a class="reference internal" href="#object.__mul__" title="object.__mul__"><code>__mul__()</code></a>, <a class="reference internal" href="#object.__rmul__" title="object.__rmul__"><code>__rmul__()</code></a> and <a class="reference internal" href="#object.__imul__" title="object.__imul__"><code>__imul__()</code></a> described below; they should not define other numerical operators. It is recommended that both mappings and sequences implement the <a class="reference internal" href="#object.__contains__" title="object.__contains__"><code>__contains__()</code></a> method to allow efficient use of the <code>in</code> operator; for mappings, <code>in</code> should search the mapping’s keys; for sequences, it should search through the values. It is further recommended that both mappings and sequences implement the <a class="reference internal" href="#object.__iter__" title="object.__iter__"><code>__iter__()</code></a> method to allow efficient iteration through the container; for mappings, <code>__iter__()</code> should iterate through the object’s keys; for sequences, it should iterate through the values.</p> <dl class="py method"> <dt class="sig sig-object py" id="object.__len__"> +<code>object.__len__(self)</code> </dt> <dd> +<p id="index-98">Called to implement the built-in function <a class="reference internal" href="../library/functions#len" title="len"><code>len()</code></a>. Should return the length of the object, an integer <code>>=</code> 0. Also, an object that doesn’t define a <a class="reference internal" href="#object.__bool__" title="object.__bool__"><code>__bool__()</code></a> method and whose <code>__len__()</code> method returns zero is considered to be false in a Boolean context.</p> <div class="impl-detail compound"> <p><strong>CPython implementation detail:</strong> In CPython, the length is required to be at most <a class="reference internal" href="../library/sys#sys.maxsize" title="sys.maxsize"><code>sys.maxsize</code></a>. If the length is larger than <code>sys.maxsize</code> some features (such as <a class="reference internal" href="../library/functions#len" title="len"><code>len()</code></a>) may raise <a class="reference internal" href="../library/exceptions#OverflowError" title="OverflowError"><code>OverflowError</code></a>. To prevent raising <code>OverflowError</code> by truth value testing, an object must define a <a class="reference internal" href="#object.__bool__" title="object.__bool__"><code>__bool__()</code></a> method.</p> </div> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__length_hint__"> +<code>object.__length_hint__(self)</code> </dt> <dd> +<p>Called to implement <a class="reference internal" href="../library/operator#operator.length_hint" title="operator.length_hint"><code>operator.length_hint()</code></a>. Should return an estimated length for the object (which may be greater or less than the actual length). The length must be an integer <code>>=</code> 0. The return value may also be <a class="reference internal" href="../library/constants#NotImplemented" title="NotImplemented"><code>NotImplemented</code></a>, which is treated the same as if the <code>__length_hint__</code> method didn’t exist at all. This method is purely an optimization and is never required for correctness.</p> <div class="versionadded"> <p><span class="versionmodified added">New in version 3.4.</span></p> </div> </dd> +</dl> <div class="admonition note" id="index-99"> <p class="admonition-title">Note</p> <p>Slicing is done exclusively with the following three methods. A call like</p> <pre data-language="python">a[1:2] = b +</pre> <p>is translated to</p> <pre data-language="python">a[slice(1, 2, None)] = b +</pre> <p>and so forth. Missing slice items are always filled in with <code>None</code>.</p> </div> <dl class="py method"> <dt class="sig sig-object py" id="object.__getitem__"> +<code>object.__getitem__(self, key)</code> </dt> <dd> +<p>Called to implement evaluation of <code>self[key]</code>. For <a class="reference internal" href="../glossary#term-sequence"><span class="xref std std-term">sequence</span></a> types, the accepted keys should be integers. Optionally, they may support <a class="reference internal" href="../library/functions#slice" title="slice"><code>slice</code></a> objects as well. Negative index support is also optional. If <em>key</em> is of an inappropriate type, <a class="reference internal" href="../library/exceptions#TypeError" title="TypeError"><code>TypeError</code></a> may be raised; if <em>key</em> is a value outside the set of indexes for the sequence (after any special interpretation of negative values), <a class="reference internal" href="../library/exceptions#IndexError" title="IndexError"><code>IndexError</code></a> should be raised. For <a class="reference internal" href="../glossary#term-mapping"><span class="xref std std-term">mapping</span></a> types, if <em>key</em> is missing (not in the container), <a class="reference internal" href="../library/exceptions#KeyError" title="KeyError"><code>KeyError</code></a> should be raised.</p> <div class="admonition note"> <p class="admonition-title">Note</p> <p><a class="reference internal" href="compound_stmts#for"><code>for</code></a> loops expect that an <a class="reference internal" href="../library/exceptions#IndexError" title="IndexError"><code>IndexError</code></a> will be raised for illegal indexes to allow proper detection of the end of the sequence.</p> </div> <div class="admonition note"> <p class="admonition-title">Note</p> <p>When <a class="reference internal" href="expressions#subscriptions"><span class="std std-ref">subscripting</span></a> a <em>class</em>, the special class method <a class="reference internal" href="#object.__class_getitem__" title="object.__class_getitem__"><code>__class_getitem__()</code></a> may be called instead of <code>__getitem__()</code>. See <a class="reference internal" href="#classgetitem-versus-getitem"><span class="std std-ref">__class_getitem__ versus __getitem__</span></a> for more details.</p> </div> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__setitem__"> +<code>object.__setitem__(self, key, value)</code> </dt> <dd> +<p>Called to implement assignment to <code>self[key]</code>. Same note as for <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><code>__getitem__()</code></a>. This should only be implemented for mappings if the objects support changes to the values for keys, or if new keys can be added, or for sequences if elements can be replaced. The same exceptions should be raised for improper <em>key</em> values as for the <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><code>__getitem__()</code></a> method.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__delitem__"> +<code>object.__delitem__(self, key)</code> </dt> <dd> +<p>Called to implement deletion of <code>self[key]</code>. Same note as for <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><code>__getitem__()</code></a>. This should only be implemented for mappings if the objects support removal of keys, or for sequences if elements can be removed from the sequence. The same exceptions should be raised for improper <em>key</em> values as for the <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><code>__getitem__()</code></a> method.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__missing__"> +<code>object.__missing__(self, key)</code> </dt> <dd> +<p>Called by <a class="reference internal" href="../library/stdtypes#dict" title="dict"><code>dict</code></a>.<a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><code>__getitem__()</code></a> to implement <code>self[key]</code> for dict subclasses when key is not in the dictionary.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__iter__"> +<code>object.__iter__(self)</code> </dt> <dd> +<p>This method is called when an <a class="reference internal" href="../glossary#term-iterator"><span class="xref std std-term">iterator</span></a> is required for a container. This method should return a new iterator object that can iterate over all the objects in the container. For mappings, it should iterate over the keys of the container.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__reversed__"> +<code>object.__reversed__(self)</code> </dt> <dd> +<p>Called (if present) by the <a class="reference internal" href="../library/functions#reversed" title="reversed"><code>reversed()</code></a> built-in to implement reverse iteration. It should return a new iterator object that iterates over all the objects in the container in reverse order.</p> <p>If the <a class="reference internal" href="#object.__reversed__" title="object.__reversed__"><code>__reversed__()</code></a> method is not provided, the <a class="reference internal" href="../library/functions#reversed" title="reversed"><code>reversed()</code></a> built-in will fall back to using the sequence protocol (<a class="reference internal" href="#object.__len__" title="object.__len__"><code>__len__()</code></a> and <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><code>__getitem__()</code></a>). Objects that support the sequence protocol should only provide <a class="reference internal" href="#object.__reversed__" title="object.__reversed__"><code>__reversed__()</code></a> if they can provide an implementation that is more efficient than the one provided by <a class="reference internal" href="../library/functions#reversed" title="reversed"><code>reversed()</code></a>.</p> </dd> +</dl> <p>The membership test operators (<a class="reference internal" href="expressions#in"><code>in</code></a> and <a class="reference internal" href="expressions#not-in"><code>not in</code></a>) are normally implemented as an iteration through a container. However, container objects can supply the following special method with a more efficient implementation, which also does not require the object be iterable.</p> <dl class="py method"> <dt class="sig sig-object py" id="object.__contains__"> +<code>object.__contains__(self, item)</code> </dt> <dd> +<p>Called to implement membership test operators. Should return true if <em>item</em> is in <em>self</em>, false otherwise. For mapping objects, this should consider the keys of the mapping rather than the values or the key-item pairs.</p> <p>For objects that don’t define <a class="reference internal" href="#object.__contains__" title="object.__contains__"><code>__contains__()</code></a>, the membership test first tries iteration via <a class="reference internal" href="#object.__iter__" title="object.__iter__"><code>__iter__()</code></a>, then the old sequence iteration protocol via <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><code>__getitem__()</code></a>, see <a class="reference internal" href="expressions#membership-test-details"><span class="std std-ref">this section in the language reference</span></a>.</p> </dd> +</dl> </section> <section id="emulating-numeric-types"> <span id="numeric-types"></span><h3> +<span class="section-number">3.3.8. </span>Emulating numeric types</h3> <p>The following methods can be defined to emulate numeric objects. Methods corresponding to operations that are not supported by the particular kind of number implemented (e.g., bitwise operations for non-integral numbers) should be left undefined.</p> <dl class="py method"> <dt class="sig sig-object py" id="object.__add__"> +<code>object.__add__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__sub__"> +<code>object.__sub__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__mul__"> +<code>object.__mul__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__matmul__"> +<code>object.__matmul__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__truediv__"> +<code>object.__truediv__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__floordiv__"> +<code>object.__floordiv__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__mod__"> +<code>object.__mod__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__divmod__"> +<code>object.__divmod__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__pow__"> +<code>object.__pow__(self, other[, modulo])</code> </dt> <dt class="sig sig-object py" id="object.__lshift__"> +<code>object.__lshift__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__rshift__"> +<code>object.__rshift__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__and__"> +<code>object.__and__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__xor__"> +<code>object.__xor__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__or__"> +<code>object.__or__(self, other)</code> </dt> <dd> +<p id="index-100">These methods are called to implement the binary arithmetic operations (<code>+</code>, <code>-</code>, <code>*</code>, <code>@</code>, <code>/</code>, <code>//</code>, <code>%</code>, <a class="reference internal" href="../library/functions#divmod" title="divmod"><code>divmod()</code></a>, <a class="reference internal" href="../library/functions#pow" title="pow"><code>pow()</code></a>, <code>**</code>, <code><<</code>, <code>>></code>, <code>&</code>, <code>^</code>, <code>|</code>). For instance, to evaluate the expression <code>x + y</code>, where <em>x</em> is an instance of a class that has an <a class="reference internal" href="#object.__add__" title="object.__add__"><code>__add__()</code></a> method, <code>type(x).__add__(x, y)</code> is called. The <a class="reference internal" href="#object.__divmod__" title="object.__divmod__"><code>__divmod__()</code></a> method should be the equivalent to using <a class="reference internal" href="#object.__floordiv__" title="object.__floordiv__"><code>__floordiv__()</code></a> and <a class="reference internal" href="#object.__mod__" title="object.__mod__"><code>__mod__()</code></a>; it should not be related to <a class="reference internal" href="#object.__truediv__" title="object.__truediv__"><code>__truediv__()</code></a>. Note that <a class="reference internal" href="#object.__pow__" title="object.__pow__"><code>__pow__()</code></a> should be defined to accept an optional third argument if the ternary version of the built-in <a class="reference internal" href="../library/functions#pow" title="pow"><code>pow()</code></a> function is to be supported.</p> <p>If one of those methods does not support the operation with the supplied arguments, it should return <code>NotImplemented</code>.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__radd__"> +<code>object.__radd__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__rsub__"> +<code>object.__rsub__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__rmul__"> +<code>object.__rmul__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__rmatmul__"> +<code>object.__rmatmul__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__rtruediv__"> +<code>object.__rtruediv__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__rfloordiv__"> +<code>object.__rfloordiv__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__rmod__"> +<code>object.__rmod__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__rdivmod__"> +<code>object.__rdivmod__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__rpow__"> +<code>object.__rpow__(self, other[, modulo])</code> </dt> <dt class="sig sig-object py" id="object.__rlshift__"> +<code>object.__rlshift__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__rrshift__"> +<code>object.__rrshift__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__rand__"> +<code>object.__rand__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__rxor__"> +<code>object.__rxor__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__ror__"> +<code>object.__ror__(self, other)</code> </dt> <dd> +<p id="index-101">These methods are called to implement the binary arithmetic operations (<code>+</code>, <code>-</code>, <code>*</code>, <code>@</code>, <code>/</code>, <code>//</code>, <code>%</code>, <a class="reference internal" href="../library/functions#divmod" title="divmod"><code>divmod()</code></a>, <a class="reference internal" href="../library/functions#pow" title="pow"><code>pow()</code></a>, <code>**</code>, <code><<</code>, <code>>></code>, <code>&</code>, <code>^</code>, <code>|</code>) with reflected (swapped) operands. These functions are only called if the left operand does not support the corresponding operation <a class="footnote-reference brackets" href="#id19" id="id14">3</a> and the operands are of different types. <a class="footnote-reference brackets" href="#id20" id="id15">4</a> For instance, to evaluate the expression <code>x - y</code>, where <em>y</em> is an instance of a class that has an <a class="reference internal" href="#object.__rsub__" title="object.__rsub__"><code>__rsub__()</code></a> method, <code>type(y).__rsub__(y, x)</code> is called if <code>type(x).__sub__(x, y)</code> returns <em>NotImplemented</em>.</p> <p id="index-102">Note that ternary <a class="reference internal" href="../library/functions#pow" title="pow"><code>pow()</code></a> will not try calling <a class="reference internal" href="#object.__rpow__" title="object.__rpow__"><code>__rpow__()</code></a> (the coercion rules would become too complicated).</p> <div class="admonition note"> <p class="admonition-title">Note</p> <p>If the right operand’s type is a subclass of the left operand’s type and that subclass provides a different implementation of the reflected method for the operation, this method will be called before the left operand’s non-reflected method. This behavior allows subclasses to override their ancestors’ operations.</p> </div> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__iadd__"> +<code>object.__iadd__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__isub__"> +<code>object.__isub__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__imul__"> +<code>object.__imul__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__imatmul__"> +<code>object.__imatmul__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__itruediv__"> +<code>object.__itruediv__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__ifloordiv__"> +<code>object.__ifloordiv__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__imod__"> +<code>object.__imod__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__ipow__"> +<code>object.__ipow__(self, other[, modulo])</code> </dt> <dt class="sig sig-object py" id="object.__ilshift__"> +<code>object.__ilshift__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__irshift__"> +<code>object.__irshift__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__iand__"> +<code>object.__iand__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__ixor__"> +<code>object.__ixor__(self, other)</code> </dt> <dt class="sig sig-object py" id="object.__ior__"> +<code>object.__ior__(self, other)</code> </dt> <dd> +<p>These methods are called to implement the augmented arithmetic assignments (<code>+=</code>, <code>-=</code>, <code>*=</code>, <code>@=</code>, <code>/=</code>, <code>//=</code>, <code>%=</code>, <code>**=</code>, <code><<=</code>, <code>>>=</code>, <code>&=</code>, <code>^=</code>, <code>|=</code>). These methods should attempt to do the operation in-place (modifying <em>self</em>) and return the result (which could be, but does not have to be, <em>self</em>). If a specific method is not defined, the augmented assignment falls back to the normal methods. For instance, if <em>x</em> is an instance of a class with an <a class="reference internal" href="#object.__iadd__" title="object.__iadd__"><code>__iadd__()</code></a> method, <code>x += y</code> is equivalent to <code>x = x.__iadd__(y)</code> . Otherwise, <code>x.__add__(y)</code> and <code>y.__radd__(x)</code> are considered, as with the evaluation of <code>x + y</code>. In certain situations, augmented assignment can result in unexpected errors (see <a class="reference internal" href="../faq/programming#faq-augmented-assignment-tuple-error"><span class="std std-ref">Why does a_tuple[i] += [‘item’] raise an exception when the addition works?</span></a>), but this behavior is in fact part of the data model.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__neg__"> +<code>object.__neg__(self)</code> </dt> <dt class="sig sig-object py" id="object.__pos__"> +<code>object.__pos__(self)</code> </dt> <dt class="sig sig-object py" id="object.__abs__"> +<code>object.__abs__(self)</code> </dt> <dt class="sig sig-object py" id="object.__invert__"> +<code>object.__invert__(self)</code> </dt> <dd> +<p id="index-103">Called to implement the unary arithmetic operations (<code>-</code>, <code>+</code>, <a class="reference internal" href="../library/functions#abs" title="abs"><code>abs()</code></a> and <code>~</code>).</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__complex__"> +<code>object.__complex__(self)</code> </dt> <dt class="sig sig-object py" id="object.__int__"> +<code>object.__int__(self)</code> </dt> <dt class="sig sig-object py" id="object.__float__"> +<code>object.__float__(self)</code> </dt> <dd> +<p id="index-104">Called to implement the built-in functions <a class="reference internal" href="../library/functions#complex" title="complex"><code>complex()</code></a>, <a class="reference internal" href="../library/functions#int" title="int"><code>int()</code></a> and <a class="reference internal" href="../library/functions#float" title="float"><code>float()</code></a>. Should return a value of the appropriate type.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__index__"> +<code>object.__index__(self)</code> </dt> <dd> +<p>Called to implement <a class="reference internal" href="../library/operator#operator.index" title="operator.index"><code>operator.index()</code></a>, and whenever Python needs to losslessly convert the numeric object to an integer object (such as in slicing, or in the built-in <a class="reference internal" href="../library/functions#bin" title="bin"><code>bin()</code></a>, <a class="reference internal" href="../library/functions#hex" title="hex"><code>hex()</code></a> and <a class="reference internal" href="../library/functions#oct" title="oct"><code>oct()</code></a> functions). Presence of this method indicates that the numeric object is an integer type. Must return an integer.</p> <p>If <a class="reference internal" href="#object.__int__" title="object.__int__"><code>__int__()</code></a>, <a class="reference internal" href="#object.__float__" title="object.__float__"><code>__float__()</code></a> and <a class="reference internal" href="#object.__complex__" title="object.__complex__"><code>__complex__()</code></a> are not defined then corresponding built-in functions <a class="reference internal" href="../library/functions#int" title="int"><code>int()</code></a>, <a class="reference internal" href="../library/functions#float" title="float"><code>float()</code></a> and <a class="reference internal" href="../library/functions#complex" title="complex"><code>complex()</code></a> fall back to <a class="reference internal" href="#object.__index__" title="object.__index__"><code>__index__()</code></a>.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__round__"> +<code>object.__round__(self[, ndigits])</code> </dt> <dt class="sig sig-object py" id="object.__trunc__"> +<code>object.__trunc__(self)</code> </dt> <dt class="sig sig-object py" id="object.__floor__"> +<code>object.__floor__(self)</code> </dt> <dt class="sig sig-object py" id="object.__ceil__"> +<code>object.__ceil__(self)</code> </dt> <dd> +<p id="index-105">Called to implement the built-in function <a class="reference internal" href="../library/functions#round" title="round"><code>round()</code></a> and <a class="reference internal" href="../library/math#module-math" title="math: Mathematical functions (sin() etc.)."><code>math</code></a> functions <a class="reference internal" href="../library/math#math.trunc" title="math.trunc"><code>trunc()</code></a>, <a class="reference internal" href="../library/math#math.floor" title="math.floor"><code>floor()</code></a> and <a class="reference internal" href="../library/math#math.ceil" title="math.ceil"><code>ceil()</code></a>. Unless <em>ndigits</em> is passed to <code>__round__()</code> all these methods should return the value of the object truncated to an <a class="reference internal" href="../library/numbers#numbers.Integral" title="numbers.Integral"><code>Integral</code></a> (typically an <a class="reference internal" href="../library/functions#int" title="int"><code>int</code></a>).</p> <p>The built-in function <a class="reference internal" href="../library/functions#int" title="int"><code>int()</code></a> falls back to <a class="reference internal" href="#object.__trunc__" title="object.__trunc__"><code>__trunc__()</code></a> if neither <a class="reference internal" href="#object.__int__" title="object.__int__"><code>__int__()</code></a> nor <a class="reference internal" href="#object.__index__" title="object.__index__"><code>__index__()</code></a> is defined.</p> <div class="versionchanged"> <p><span class="versionmodified changed">Changed in version 3.11: </span>The delegation of <a class="reference internal" href="../library/functions#int" title="int"><code>int()</code></a> to <a class="reference internal" href="#object.__trunc__" title="object.__trunc__"><code>__trunc__()</code></a> is deprecated.</p> </div> </dd> +</dl> </section> <section id="with-statement-context-managers"> <span id="context-managers"></span><h3> +<span class="section-number">3.3.9. </span>With Statement Context Managers</h3> <p>A <em class="dfn">context manager</em> is an object that defines the runtime context to be established when executing a <a class="reference internal" href="compound_stmts#with"><code>with</code></a> statement. The context manager handles the entry into, and the exit from, the desired runtime context for the execution of the block of code. Context managers are normally invoked using the <code>with</code> statement (described in section <a class="reference internal" href="compound_stmts#with"><span class="std std-ref">The with statement</span></a>), but can also be used by directly invoking their methods.</p> <p id="index-106">Typical uses of context managers include saving and restoring various kinds of global state, locking and unlocking resources, closing opened files, etc.</p> <p>For more information on context managers, see <a class="reference internal" href="../library/stdtypes#typecontextmanager"><span class="std std-ref">Context Manager Types</span></a>.</p> <dl class="py method"> <dt class="sig sig-object py" id="object.__enter__"> +<code>object.__enter__(self)</code> </dt> <dd> +<p>Enter the runtime context related to this object. The <a class="reference internal" href="compound_stmts#with"><code>with</code></a> statement will bind this method’s return value to the target(s) specified in the <code>as</code> clause of the statement, if any.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__exit__"> +<code>object.__exit__(self, exc_type, exc_value, traceback)</code> </dt> <dd> +<p>Exit the runtime context related to this object. The parameters describe the exception that caused the context to be exited. If the context was exited without an exception, all three arguments will be <a class="reference internal" href="../library/constants#None" title="None"><code>None</code></a>.</p> <p>If an exception is supplied, and the method wishes to suppress the exception (i.e., prevent it from being propagated), it should return a true value. Otherwise, the exception will be processed normally upon exit from this method.</p> <p>Note that <a class="reference internal" href="#object.__exit__" title="object.__exit__"><code>__exit__()</code></a> methods should not reraise the passed-in exception; this is the caller’s responsibility.</p> </dd> +</dl> <div class="admonition seealso"> <p class="admonition-title">See also</p> <dl class="simple"> <dt> +<span class="target" id="index-107"></span><a class="pep reference external" href="https://peps.python.org/pep-0343/"><strong>PEP 343</strong></a> - The “with” statement</dt> +<dd> +<p>The specification, background, and examples for the Python <a class="reference internal" href="compound_stmts#with"><code>with</code></a> statement.</p> </dd> </dl> </div> </section> <section id="customizing-positional-arguments-in-class-pattern-matching"> <span id="class-pattern-matching"></span><h3> +<span class="section-number">3.3.10. </span>Customizing positional arguments in class pattern matching</h3> <p>When using a class name in a pattern, positional arguments in the pattern are not allowed by default, i.e. <code>case MyClass(x, y)</code> is typically invalid without special support in <code>MyClass</code>. To be able to use that kind of pattern, the class needs to define a <em>__match_args__</em> attribute.</p> <dl class="py data"> <dt class="sig sig-object py" id="object.__match_args__"> +<code>object.__match_args__</code> </dt> <dd> +<p>This class variable can be assigned a tuple of strings. When this class is used in a class pattern with positional arguments, each positional argument will be converted into a keyword argument, using the corresponding value in <em>__match_args__</em> as the keyword. The absence of this attribute is equivalent to setting it to <code>()</code>.</p> </dd> +</dl> <p>For example, if <code>MyClass.__match_args__</code> is <code>("left", "center", "right")</code> that means that <code>case MyClass(x, y)</code> is equivalent to <code>case MyClass(left=x, center=y)</code>. Note that the number of arguments in the pattern must be smaller than or equal to the number of elements in <em>__match_args__</em>; if it is larger, the pattern match attempt will raise a <a class="reference internal" href="../library/exceptions#TypeError" title="TypeError"><code>TypeError</code></a>.</p> <div class="versionadded"> <p><span class="versionmodified added">New in version 3.10.</span></p> </div> <div class="admonition seealso"> <p class="admonition-title">See also</p> <dl class="simple"> <dt> +<span class="target" id="index-108"></span><a class="pep reference external" href="https://peps.python.org/pep-0634/"><strong>PEP 634</strong></a> - Structural Pattern Matching</dt> +<dd> +<p>The specification for the Python <code>match</code> statement.</p> </dd> </dl> </div> </section> <section id="emulating-buffer-types"> <span id="python-buffer-protocol"></span><h3> +<span class="section-number">3.3.11. </span>Emulating buffer types</h3> <p>The <a class="reference internal" href="../c-api/buffer#bufferobjects"><span class="std std-ref">buffer protocol</span></a> provides a way for Python objects to expose efficient access to a low-level memory array. This protocol is implemented by builtin types such as <a class="reference internal" href="../library/stdtypes#bytes" title="bytes"><code>bytes</code></a> and <a class="reference internal" href="../library/stdtypes#memoryview" title="memoryview"><code>memoryview</code></a>, and third-party libraries may define additional buffer types.</p> <p>While buffer types are usually implemented in C, it is also possible to implement the protocol in Python.</p> <dl class="py method"> <dt class="sig sig-object py" id="object.__buffer__"> +<code>object.__buffer__(self, flags)</code> </dt> <dd> +<p>Called when a buffer is requested from <em>self</em> (for example, by the <a class="reference internal" href="../library/stdtypes#memoryview" title="memoryview"><code>memoryview</code></a> constructor). The <em>flags</em> argument is an integer representing the kind of buffer requested, affecting for example whether the returned buffer is read-only or writable. <a class="reference internal" href="../library/inspect#inspect.BufferFlags" title="inspect.BufferFlags"><code>inspect.BufferFlags</code></a> provides a convenient way to interpret the flags. The method must return a <a class="reference internal" href="../library/stdtypes#memoryview" title="memoryview"><code>memoryview</code></a> object.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__release_buffer__"> +<code>object.__release_buffer__(self, buffer)</code> </dt> <dd> +<p>Called when a buffer is no longer needed. The <em>buffer</em> argument is a <a class="reference internal" href="../library/stdtypes#memoryview" title="memoryview"><code>memoryview</code></a> object that was previously returned by <a class="reference internal" href="#object.__buffer__" title="object.__buffer__"><code>__buffer__()</code></a>. The method must release any resources associated with the buffer. This method should return <code>None</code>. Buffer objects that do not need to perform any cleanup are not required to implement this method.</p> </dd> +</dl> <div class="versionadded"> <p><span class="versionmodified added">New in version 3.12.</span></p> </div> <div class="admonition seealso"> <p class="admonition-title">See also</p> <dl class="simple"> <dt> +<span class="target" id="index-109"></span><a class="pep reference external" href="https://peps.python.org/pep-0688/"><strong>PEP 688</strong></a> - Making the buffer protocol accessible in Python</dt> +<dd> +<p>Introduces the Python <code>__buffer__</code> and <code>__release_buffer__</code> methods.</p> </dd> <dt> +<code></code> <a class="reference internal" href="../library/collections.abc#collections.abc.Buffer" title="collections.abc.Buffer"><code>collections.abc.Buffer</code></a> +</dt> +<dd> +<p>ABC for buffer types.</p> </dd> </dl> </div> </section> <section id="special-method-lookup"> <span id="special-lookup"></span><h3> +<span class="section-number">3.3.12. </span>Special method lookup</h3> <p>For custom classes, implicit invocations of special methods are only guaranteed to work correctly if defined on an object’s type, not in the object’s instance dictionary. That behaviour is the reason why the following code raises an exception:</p> <pre data-language="python">>>> class C: +... pass +... +>>> c = C() +>>> c.__len__ = lambda: 5 +>>> len(c) +Traceback (most recent call last): + File "<stdin>", line 1, in <module> +TypeError: object of type 'C' has no len() +</pre> <p>The rationale behind this behaviour lies with a number of special methods such as <a class="reference internal" href="#object.__hash__" title="object.__hash__"><code>__hash__()</code></a> and <a class="reference internal" href="#object.__repr__" title="object.__repr__"><code>__repr__()</code></a> that are implemented by all objects, including type objects. If the implicit lookup of these methods used the conventional lookup process, they would fail when invoked on the type object itself:</p> <pre data-language="python">>>> 1 .__hash__() == hash(1) +True +>>> int.__hash__() == hash(int) +Traceback (most recent call last): + File "<stdin>", line 1, in <module> +TypeError: descriptor '__hash__' of 'int' object needs an argument +</pre> <p>Incorrectly attempting to invoke an unbound method of a class in this way is sometimes referred to as ‘metaclass confusion’, and is avoided by bypassing the instance when looking up special methods:</p> <pre data-language="python">>>> type(1).__hash__(1) == hash(1) +True +>>> type(int).__hash__(int) == hash(int) +True +</pre> <p>In addition to bypassing any instance attributes in the interest of correctness, implicit special method lookup generally also bypasses the <a class="reference internal" href="#object.__getattribute__" title="object.__getattribute__"><code>__getattribute__()</code></a> method even of the object’s metaclass:</p> <pre data-language="python">>>> class Meta(type): +... def __getattribute__(*args): +... print("Metaclass getattribute invoked") +... return type.__getattribute__(*args) +... +>>> class C(object, metaclass=Meta): +... def __len__(self): +... return 10 +... def __getattribute__(*args): +... print("Class getattribute invoked") +... return object.__getattribute__(*args) +... +>>> c = C() +>>> c.__len__() # Explicit lookup via instance +Class getattribute invoked +10 +>>> type(c).__len__(c) # Explicit lookup via type +Metaclass getattribute invoked +10 +>>> len(c) # Implicit lookup +10 +</pre> <p>Bypassing the <a class="reference internal" href="#object.__getattribute__" title="object.__getattribute__"><code>__getattribute__()</code></a> machinery in this fashion provides significant scope for speed optimisations within the interpreter, at the cost of some flexibility in the handling of special methods (the special method <em>must</em> be set on the class object itself in order to be consistently invoked by the interpreter).</p> </section> </section> <section id="coroutines"> <span id="index-110"></span><h2> +<span class="section-number">3.4. </span>Coroutines</h2> <section id="awaitable-objects"> <h3> +<span class="section-number">3.4.1. </span>Awaitable Objects</h3> <p>An <a class="reference internal" href="../glossary#term-awaitable"><span class="xref std std-term">awaitable</span></a> object generally implements an <a class="reference internal" href="#object.__await__" title="object.__await__"><code>__await__()</code></a> method. <a class="reference internal" href="../glossary#term-coroutine"><span class="xref std std-term">Coroutine objects</span></a> returned from <a class="reference internal" href="compound_stmts#async-def"><code>async def</code></a> functions are awaitable.</p> <div class="admonition note"> <p class="admonition-title">Note</p> <p>The <a class="reference internal" href="../glossary#term-generator-iterator"><span class="xref std std-term">generator iterator</span></a> objects returned from generators decorated with <a class="reference internal" href="../library/types#types.coroutine" title="types.coroutine"><code>types.coroutine()</code></a> are also awaitable, but they do not implement <a class="reference internal" href="#object.__await__" title="object.__await__"><code>__await__()</code></a>.</p> </div> <dl class="py method"> <dt class="sig sig-object py" id="object.__await__"> +<code>object.__await__(self)</code> </dt> <dd> +<p>Must return an <a class="reference internal" href="../glossary#term-iterator"><span class="xref std std-term">iterator</span></a>. Should be used to implement <a class="reference internal" href="../glossary#term-awaitable"><span class="xref std std-term">awaitable</span></a> objects. For instance, <a class="reference internal" href="../library/asyncio-future#asyncio.Future" title="asyncio.Future"><code>asyncio.Future</code></a> implements this method to be compatible with the <a class="reference internal" href="expressions#await"><code>await</code></a> expression.</p> <div class="admonition note"> <p class="admonition-title">Note</p> <p>The language doesn’t place any restriction on the type or value of the objects yielded by the iterator returned by <code>__await__</code>, as this is specific to the implementation of the asynchronous execution framework (e.g. <a class="reference internal" href="../library/asyncio#module-asyncio" title="asyncio: Asynchronous I/O."><code>asyncio</code></a>) that will be managing the <a class="reference internal" href="../glossary#term-awaitable"><span class="xref std std-term">awaitable</span></a> object.</p> </div> </dd> +</dl> <div class="versionadded"> <p><span class="versionmodified added">New in version 3.5.</span></p> </div> <div class="admonition seealso"> <p class="admonition-title">See also</p> <p><span class="target" id="index-111"></span><a class="pep reference external" href="https://peps.python.org/pep-0492/"><strong>PEP 492</strong></a> for additional information about awaitable objects.</p> </div> </section> <section id="coroutine-objects"> <span id="id16"></span><h3> +<span class="section-number">3.4.2. </span>Coroutine Objects</h3> <p><a class="reference internal" href="../glossary#term-coroutine"><span class="xref std std-term">Coroutine objects</span></a> are <a class="reference internal" href="../glossary#term-awaitable"><span class="xref std std-term">awaitable</span></a> objects. A coroutine’s execution can be controlled by calling <a class="reference internal" href="#object.__await__" title="object.__await__"><code>__await__()</code></a> and iterating over the result. When the coroutine has finished executing and returns, the iterator raises <a class="reference internal" href="../library/exceptions#StopIteration" title="StopIteration"><code>StopIteration</code></a>, and the exception’s <a class="reference internal" href="../library/exceptions#StopIteration.value" title="StopIteration.value"><code>value</code></a> attribute holds the return value. If the coroutine raises an exception, it is propagated by the iterator. Coroutines should not directly raise unhandled <a class="reference internal" href="../library/exceptions#StopIteration" title="StopIteration"><code>StopIteration</code></a> exceptions.</p> <p>Coroutines also have the methods listed below, which are analogous to those of generators (see <a class="reference internal" href="expressions#generator-methods"><span class="std std-ref">Generator-iterator methods</span></a>). However, unlike generators, coroutines do not directly support iteration.</p> <div class="versionchanged"> <p><span class="versionmodified changed">Changed in version 3.5.2: </span>It is a <a class="reference internal" href="../library/exceptions#RuntimeError" title="RuntimeError"><code>RuntimeError</code></a> to await on a coroutine more than once.</p> </div> <dl class="py method"> <dt class="sig sig-object py" id="coroutine.send"> +<code>coroutine.send(value)</code> </dt> <dd> +<p>Starts or resumes execution of the coroutine. If <em>value</em> is <code>None</code>, this is equivalent to advancing the iterator returned by <a class="reference internal" href="#object.__await__" title="object.__await__"><code>__await__()</code></a>. If <em>value</em> is not <code>None</code>, this method delegates to the <a class="reference internal" href="expressions#generator.send" title="generator.send"><code>send()</code></a> method of the iterator that caused the coroutine to suspend. The result (return value, <a class="reference internal" href="../library/exceptions#StopIteration" title="StopIteration"><code>StopIteration</code></a>, or other exception) is the same as when iterating over the <code>__await__()</code> return value, described above.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="coroutine.throw"> +<code>coroutine.throw(value)</code> </dt> <dt class="sig sig-object py"> <span class="sig-prename descclassname">coroutine.</span><span class="sig-name descname">throw</span><span class="sig-paren">(</span><em class="sig-param"><span class="n">type</span></em><span class="optional">[</span>, <em class="sig-param"><span class="n">value</span></em><span class="optional">[</span>, <em class="sig-param"><span class="n">traceback</span></em><span class="optional">]</span><span class="optional">]</span><span class="sig-paren">)</span> +</dt> <dd> +<p>Raises the specified exception in the coroutine. This method delegates to the <a class="reference internal" href="expressions#generator.throw" title="generator.throw"><code>throw()</code></a> method of the iterator that caused the coroutine to suspend, if it has such a method. Otherwise, the exception is raised at the suspension point. The result (return value, <a class="reference internal" href="../library/exceptions#StopIteration" title="StopIteration"><code>StopIteration</code></a>, or other exception) is the same as when iterating over the <a class="reference internal" href="#object.__await__" title="object.__await__"><code>__await__()</code></a> return value, described above. If the exception is not caught in the coroutine, it propagates back to the caller.</p> <div class="versionchanged"> <p><span class="versionmodified changed">Changed in version 3.12: </span>The second signature (type[, value[, traceback]]) is deprecated and may be removed in a future version of Python.</p> </div> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="coroutine.close"> +<code>coroutine.close()</code> </dt> <dd> +<p>Causes the coroutine to clean itself up and exit. If the coroutine is suspended, this method first delegates to the <a class="reference internal" href="expressions#generator.close" title="generator.close"><code>close()</code></a> method of the iterator that caused the coroutine to suspend, if it has such a method. Then it raises <a class="reference internal" href="../library/exceptions#GeneratorExit" title="GeneratorExit"><code>GeneratorExit</code></a> at the suspension point, causing the coroutine to immediately clean itself up. Finally, the coroutine is marked as having finished executing, even if it was never started.</p> <p>Coroutine objects are automatically closed using the above process when they are about to be destroyed.</p> </dd> +</dl> </section> <section id="asynchronous-iterators"> <span id="async-iterators"></span><h3> +<span class="section-number">3.4.3. </span>Asynchronous Iterators</h3> <p>An <em>asynchronous iterator</em> can call asynchronous code in its <code>__anext__</code> method.</p> <p>Asynchronous iterators can be used in an <a class="reference internal" href="compound_stmts#async-for"><code>async for</code></a> statement.</p> <dl class="py method"> <dt class="sig sig-object py" id="object.__aiter__"> +<code>object.__aiter__(self)</code> </dt> <dd> +<p>Must return an <em>asynchronous iterator</em> object.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__anext__"> +<code>object.__anext__(self)</code> </dt> <dd> +<p>Must return an <em>awaitable</em> resulting in a next value of the iterator. Should raise a <a class="reference internal" href="../library/exceptions#StopAsyncIteration" title="StopAsyncIteration"><code>StopAsyncIteration</code></a> error when the iteration is over.</p> </dd> +</dl> <p>An example of an asynchronous iterable object:</p> <pre data-language="python">class Reader: + async def readline(self): + ... + + def __aiter__(self): + return self + + async def __anext__(self): + val = await self.readline() + if val == b'': + raise StopAsyncIteration + return val +</pre> <div class="versionadded"> <p><span class="versionmodified added">New in version 3.5.</span></p> </div> <div class="versionchanged"> <p><span class="versionmodified changed">Changed in version 3.7: </span>Prior to Python 3.7, <a class="reference internal" href="#object.__aiter__" title="object.__aiter__"><code>__aiter__()</code></a> could return an <em>awaitable</em> that would resolve to an <a class="reference internal" href="../glossary#term-asynchronous-iterator"><span class="xref std std-term">asynchronous iterator</span></a>.</p> <p>Starting with Python 3.7, <a class="reference internal" href="#object.__aiter__" title="object.__aiter__"><code>__aiter__()</code></a> must return an asynchronous iterator object. Returning anything else will result in a <a class="reference internal" href="../library/exceptions#TypeError" title="TypeError"><code>TypeError</code></a> error.</p> </div> </section> <section id="asynchronous-context-managers"> <span id="async-context-managers"></span><h3> +<span class="section-number">3.4.4. </span>Asynchronous Context Managers</h3> <p>An <em>asynchronous context manager</em> is a <em>context manager</em> that is able to suspend execution in its <code>__aenter__</code> and <code>__aexit__</code> methods.</p> <p>Asynchronous context managers can be used in an <a class="reference internal" href="compound_stmts#async-with"><code>async with</code></a> statement.</p> <dl class="py method"> <dt class="sig sig-object py" id="object.__aenter__"> +<code>object.__aenter__(self)</code> </dt> <dd> +<p>Semantically similar to <a class="reference internal" href="#object.__enter__" title="object.__enter__"><code>__enter__()</code></a>, the only difference being that it must return an <em>awaitable</em>.</p> </dd> +</dl> <dl class="py method"> <dt class="sig sig-object py" id="object.__aexit__"> +<code>object.__aexit__(self, exc_type, exc_value, traceback)</code> </dt> <dd> +<p>Semantically similar to <a class="reference internal" href="#object.__exit__" title="object.__exit__"><code>__exit__()</code></a>, the only difference being that it must return an <em>awaitable</em>.</p> </dd> +</dl> <p>An example of an asynchronous context manager class:</p> <pre data-language="python">class AsyncContextManager: + async def __aenter__(self): + await log('entering context') + + async def __aexit__(self, exc_type, exc, tb): + await log('exiting context') +</pre> <div class="versionadded"> <p><span class="versionmodified added">New in version 3.5.</span></p> </div> <h4 class="rubric">Footnotes</h4> <dl class="footnote brackets"> <dt class="label" id="id17"> +<code>1</code> </dt> <dd> +<p>It <em>is</em> possible in some cases to change an object’s type, under certain controlled conditions. It generally isn’t a good idea though, since it can lead to some very strange behaviour if it is handled incorrectly.</p> </dd> <dt class="label" id="id18"> +<code>2</code> </dt> <dd> +<p>The <a class="reference internal" href="#object.__hash__" title="object.__hash__"><code>__hash__()</code></a>, <a class="reference internal" href="#object.__iter__" title="object.__iter__"><code>__iter__()</code></a>, <a class="reference internal" href="#object.__reversed__" title="object.__reversed__"><code>__reversed__()</code></a>, and <a class="reference internal" href="#object.__contains__" title="object.__contains__"><code>__contains__()</code></a> methods have special handling for this; others will still raise a <a class="reference internal" href="../library/exceptions#TypeError" title="TypeError"><code>TypeError</code></a>, but may do so by relying on the behavior that <code>None</code> is not callable.</p> </dd> <dt class="label" id="id19"> +<code>3</code> </dt> <dd> +<p>“Does not support” here means that the class has no such method, or the method returns <code>NotImplemented</code>. Do not set the method to <code>None</code> if you want to force fallback to the right operand’s reflected method—that will instead have the opposite effect of explicitly <em>blocking</em> such fallback.</p> </dd> <dt class="label" id="id20"> +<code>4</code> </dt> <dd> +<p>For operands of the same type, it is assumed that if the non-reflected method – such as <a class="reference internal" href="#object.__add__" title="object.__add__"><code>__add__()</code></a> – fails then the overall operation is not supported, which is why the reflected method is not called.</p> </dd> </dl> </section> </section> <div class="_attribution"> + <p class="_attribution-p"> + © 2001–2023 Python Software Foundation<br>Licensed under the PSF License.<br> + <a href="https://docs.python.org/3.12/reference/datamodel.html" class="_attribution-link">https://docs.python.org/3.12/reference/datamodel.html</a> + </p> +</div> |