29.12. inspect — Inspect live objects

Python 3.4.4

29.12. inspect — Inspect live objects

Source code: Lib/inspect.py


The inspect module provides several useful functions to help get information about live objects such as modules, classes, methods, functions, tracebacks, frame objects, and code objects. For example, it can help you examine the contents of a class, retrieve the source code of a method, extract and format the argument list for a function, or get all the information you need to display a detailed traceback.

There are four main kinds of services provided by this module: type checking, getting source code, inspecting classes and functions, and examining the interpreter stack.

29.12.1. Types and members

The getmembers() function retrieves the members of an object such as a class or module. The sixteen functions whose names begin with “is” are mainly provided as convenient choices for the second argument to getmembers(). They also help you determine when you can expect to find the following special attributes:

Type Attribute Description
module __doc__ documentation string
  __file__ filename (missing for built-in modules)
class __doc__ documentation string
  __name__ name with which this class was defined
  __qualname__ qualified name
  __module__ name of module in which this class was defined
method __doc__ documentation string
  __name__ name with which this method was defined
  __qualname__ qualified name
  __func__ function object containing implementation of method
  __self__ instance to which this method is bound, or None
function __doc__ documentation string
  __name__ name with which this function was defined
  __qualname__ qualified name
  __code__ code object containing compiled function bytecode
  __defaults__ tuple of any default values for positional or keyword parameters
  __kwdefaults__ mapping of any default values for keyword-only parameters
  __globals__ global namespace in which this function was defined
traceback tb_frame frame object at this level
  tb_lasti index of last attempted instruction in bytecode
  tb_lineno current line number in Python source code
  tb_next next inner traceback object (called by this level)
frame f_back next outer frame object (this frame’s caller)
  f_builtins builtins namespace seen by this frame
  f_code code object being executed in this frame
  f_globals global namespace seen by this frame
  f_lasti index of last attempted instruction in bytecode
  f_lineno current line number in Python source code
  f_locals local namespace seen by this frame
  f_restricted 0 or 1 if frame is in restricted execution mode
  f_trace tracing function for this frame, or None
code co_argcount number of arguments (not including * or ** args)
  co_code string of raw compiled bytecode
  co_consts tuple of constants used in the bytecode
  co_filename name of file in which this code object was created
  co_firstlineno number of first line in Python source code
  co_flags bitmap: 1=optimized | 2=newlocals | 4=*arg | 8=**arg
  co_lnotab encoded mapping of line numbers to bytecode indices
  co_name name with which this code object was defined
  co_names tuple of names of local variables
  co_nlocals number of local variables
  co_stacksize virtual machine stack space required
  co_varnames tuple of names of arguments and local variables
builtin __doc__ documentation string
  __name__ original name of this function or method
  __qualname__ qualified name
  __self__ instance to which a method is bound, or None
inspect.getmembers(object[, predicate])

Return all the members of an object in a list of (name, value) pairs sorted by name. If the optional predicate argument is supplied, only members for which the predicate returns a true value are included.

Note

getmembers() will only return class attributes defined in the metaclass when the argument is a class and those attributes have been listed in the metaclass’ custom __dir__().

inspect.getmoduleinfo(path)

Returns a named tuple ModuleInfo(name, suffix, mode, module_type) of values that describe how Python will interpret the file identified by path if it is a module, or None if it would not be identified as a module. In that tuple, name is the name of the module without the name of any enclosing package, suffix is the trailing part of the file name (which may not be a dot-delimited extension), mode is the open() mode that would be used ('r' or 'rb'), and module_type is an integer giving the type of the module. module_type will have a value which can be compared to the constants defined in the imp module; see the documentation for that module for more information on module types.

Deprecated since version 3.3: You may check the file path’s suffix against the supported suffixes listed in importlib.machinery to infer the same information.

inspect.getmodulename(path)

Return the name of the module named by the file path, without including the names of enclosing packages. The file extension is checked against all of the entries in importlib.machinery.all_suffixes(). If it matches, the final path component is returned with the extension removed. Otherwise, None is returned.

Note that this function only returns a meaningful name for actual Python modules - paths that potentially refer to Python packages will still return None.

Changed in version 3.3: This function is now based directly on importlib rather than the deprecated getmoduleinfo().

inspect.ismodule(object)

Return true if the object is a module.

inspect.isclass(object)

Return true if the object is a class, whether built-in or created in Python code.

inspect.ismethod(object)

Return true if the object is a bound method written in Python.

inspect.isfunction(object)

Return true if the object is a Python function, which includes functions created by a lambda expression.

inspect.isgeneratorfunction(object)

Return true if the object is a Python generator function.

inspect.isgenerator(object)

Return true if the object is a generator.

inspect.istraceback(object)

Return true if the object is a traceback.

inspect.isframe(object)

Return true if the object is a frame.

inspect.iscode(object)

Return true if the object is a code.

inspect.isbuiltin(object)

Return true if the object is a built-in function or a bound built-in method.

inspect.isroutine(object)

Return true if the object is a user-defined or built-in function or method.

inspect.isabstract(object)

Return true if the object is an abstract base class.

inspect.ismethoddescriptor(object)

Return true if the object is a method descriptor, but not if ismethod(), isclass(), isfunction() or isbuiltin() are true.

This, for example, is true of int.__add__. An object passing this test has a __get__ attribute but not a __set__ attribute, but beyond that the set of attributes varies. __name__ is usually sensible, and __doc__ often is.

Methods implemented via descriptors that also pass one of the other tests return false from the ismethoddescriptor() test, simply because the other tests promise more – you can, e.g., count on having the __func__ attribute (etc) when an object passes ismethod().

inspect.isdatadescriptor(object)

Return true if the object is a data descriptor.

Data descriptors have both a __get__ and a __set__ attribute. Examples are properties (defined in Python), getsets, and members. The latter two are defined in C and there are more specific tests available for those types, which is robust across Python implementations. Typically, data descriptors will also have __name__ and __doc__ attributes (properties, getsets, and members have both of these attributes), but this is not guaranteed.

inspect.isgetsetdescriptor(object)

Return true if the object is a getset descriptor.

CPython implementation detail: getsets are attributes defined in extension modules via PyGetSetDef structures. For Python implementations without such types, this method will always return False.

inspect.ismemberdescriptor(object)

Return true if the object is a member descriptor.

CPython implementation detail: Member descriptors are attributes defined in extension modules via PyMemberDef structures. For Python implementations without such types, this method will always return False.

29.12.2. Retrieving source code

inspect.getdoc(object)

Get the documentation string for an object, cleaned up with cleandoc().

inspect.getcomments(object)

Return in a single string any lines of comments immediately preceding the object’s source code (for a class, function, or method), or at the top of the Python source file (if the object is a module).

inspect.getfile(object)

Return the name of the (text or binary) file in which an object was defined. This will fail with a TypeError if the object is a built-in module, class, or function.

inspect.getmodule(object)

Try to guess which module an object was defined in.

inspect.getsourcefile(object)

Return the name of the Python source file in which an object was defined. This will fail with a TypeError if the object is a built-in module, class, or function.

inspect.getsourcelines(object)

Return a list of source lines and starting line number for an object. The argument may be a module, class, method, function, traceback, frame, or code object. The source code is returned as a list of the lines corresponding to the object and the line number indicates where in the original source file the first line of code was found. An OSError is raised if the source code cannot be retrieved.

Changed in version 3.3: OSError is raised instead of IOError, now an alias of the former.

inspect.getsource(object)

Return the text of the source code for an object. The argument may be a module, class, method, function, traceback, frame, or code object. The source code is returned as a single string. An OSError is raised if the source code cannot be retrieved.

Changed in version 3.3: OSError is raised instead of IOError, now an alias of the former.

inspect.cleandoc(doc)

Clean up indentation from docstrings that are indented to line up with blocks of code. Any whitespace that can be uniformly removed from the second line onwards is removed. Also, all tabs are expanded to spaces.

29.12.3. Introspecting callables with the Signature object

New in version 3.3.

The Signature object represents the call signature of a callable object and its return annotation. To retrieve a Signature object, use the signature() function.

inspect.signature(callable)

Return a Signature object for the given callable:

>>> from inspect import signature
>>> def foo(a, *, b:int, **kwargs):
...     pass

>>> sig = signature(foo)

>>> str(sig)
'(a, *, b:int, **kwargs)'

>>> str(sig.parameters['b'])
'b:int'

>>> sig.parameters['b'].annotation
<class 'int'>

Accepts a wide range of python callables, from plain functions and classes to functools.partial() objects.

Raises ValueError if no signature can be provided, and TypeError if that type of object is not supported.

Note

Some callables may not be introspectable in certain implementations of Python. For example, in CPython, some built-in functions defined in C provide no metadata about their arguments.

class inspect.Signature(parameters=None, *, return_annotation=Signature.empty)

A Signature object represents the call signature of a function and its return annotation. For each parameter accepted by the function it stores a Parameter object in its parameters collection.

The optional parameters argument is a sequence of Parameter objects, which is validated to check that there are no parameters with duplicate names, and that the parameters are in the right order, i.e. positional-only first, then positional-or-keyword, and that parameters with defaults follow parameters without defaults.

The optional return_annotation argument, can be an arbitrary Python object, is the “return” annotation of the callable.

Signature objects are immutable. Use Signature.replace() to make a modified copy.

empty

A special class-level marker to specify absence of a return annotation.

parameters

An ordered mapping of parameters’ names to the corresponding Parameter objects.

return_annotation

The “return” annotation for the callable. If the callable has no “return” annotation, this attribute is set to Signature.empty.

bind(*args, **kwargs)

Create a mapping from positional and keyword arguments to parameters. Returns BoundArguments if *args and **kwargs match the signature, or raises a TypeError.

bind_partial(*args, **kwargs)

Works the same way as Signature.bind(), but allows the omission of some required arguments (mimics functools.partial() behavior.) Returns BoundArguments, or raises a TypeError if the passed arguments do not match the signature.

replace(*[, parameters][, return_annotation])

Create a new Signature instance based on the instance replace was invoked on. It is possible to pass different parameters and/or return_annotation to override the corresponding properties of the base signature. To remove return_annotation from the copied Signature, pass in Signature.empty.

>>> def test(a, b):
...     pass
>>> sig = signature(test)
>>> new_sig = sig.replace(return_annotation="new return anno")
>>> str(new_sig)
"(a, b) -> 'new return anno'"
class inspect.Parameter(name, kind, *, default=Parameter.empty, annotation=Parameter.empty)

Parameter objects are immutable. Instead of modifying a Parameter object, you can use Parameter.replace() to create a modified copy.

empty

A special class-level marker to specify absence of default values and annotations.

name

The name of the parameter as a string. The name must be a valid Python identifier.

default

The default value for the parameter. If the parameter has no default value, this attribute is set to Parameter.empty.

annotation

The annotation for the parameter. If the parameter has no annotation, this attribute is set to Parameter.empty.

kind

Describes how argument values are bound to the parameter. Possible values (accessible via Parameter, like Parameter.KEYWORD_ONLY):

Name Meaning
POSITIONAL_ONLY

Value must be supplied as a positional argument.

Python has no explicit syntax for defining positional-only parameters, but many built-in and extension module functions (especially those that accept only one or two parameters) accept them.

POSITIONAL_OR_KEYWORD Value may be supplied as either a keyword or positional argument (this is the standard binding behaviour for functions implemented in Python.)
VAR_POSITIONAL A tuple of positional arguments that aren’t bound to any other parameter. This corresponds to a *args parameter in a Python function definition.
KEYWORD_ONLY Value must be supplied as a keyword argument. Keyword only parameters are those which appear after a * or *args entry in a Python function definition.
VAR_KEYWORD A dict of keyword arguments that aren’t bound to any other parameter. This corresponds to a **kwargs parameter in a Python function definition.

Example: print all keyword-only arguments without default values:

>>> def foo(a, b, *, c, d=10):
...     pass

>>> sig = signature(foo)
>>> for param in sig.parameters.values():
...     if (param.kind == param.KEYWORD_ONLY and
...                        param.default is param.empty):
...         print('Parameter:', param)
Parameter: c
replace(*[, name][, kind][, default][, annotation])

Create a new Parameter instance based on the instance replaced was invoked on. To override a Parameter attribute, pass the corresponding argument. To remove a default value or/and an annotation from a Parameter, pass Parameter.empty.

>>> from inspect import Parameter
>>> param = Parameter('foo', Parameter.KEYWORD_ONLY, default=42)
>>> str(param)
'foo=42'

>>> str(param.replace()) # Will create a shallow copy of 'param'
'foo=42'

>>> str(param.replace(default=Parameter.empty, annotation='spam'))
"foo:'spam'"

Changed in version 3.4: In Python 3.3 Parameter objects were allowed to have name set to None if their kind was set to POSITIONAL_ONLY. This is no longer permitted.

class inspect.BoundArguments

Result of a Signature.bind() or Signature.bind_partial() call. Holds the mapping of arguments to the function’s parameters.

arguments

An ordered, mutable mapping (collections.OrderedDict) of parameters’ names to arguments’ values. Contains only explicitly bound arguments. Changes in arguments will reflect in args and kwargs.

Should be used in conjunction with Signature.parameters for any argument processing purposes.

Note

Arguments for which Signature.bind() or Signature.bind_partial() relied on a default value are skipped. However, if needed, it is easy to include them.

>>> def foo(a, b=10):
...     pass

>>> sig = signature(foo)
>>> ba = sig.bind(5)

>>> ba.args, ba.kwargs
((5,), {})

>>> for param in sig.parameters.values():
...     if (param.name not in ba.arguments
...             and param.default is not param.empty):
...         ba.arguments[param.name] = param.default

>>> ba.args, ba.kwargs
((5, 10), {})
args

A tuple of positional arguments values. Dynamically computed from the arguments attribute.

kwargs

A dict of keyword arguments values. Dynamically computed from the arguments attribute.

signature

A reference to the parent Signature object.

The args and kwargs properties can be used to invoke functions:

def test(a, *, b):
   ...

sig = signature(test)
ba = sig.bind(10, b=20)
test(*ba.args, **ba.kwargs)

See also

PEP 362 - Function Signature Object.
The detailed specification, implementation details and examples.

29.12.4. Classes and functions

inspect.getclasstree(classes, unique=False)

Arrange the given list of classes into a hierarchy of nested lists. Where a nested list appears, it contains classes derived from the class whose entry immediately precedes the list. Each entry is a 2-tuple containing a class and a tuple of its base classes. If the unique argument is true, exactly one entry appears in the returned structure for each class in the given list. Otherwise, classes using multiple inheritance and their descendants will appear multiple times.

inspect.getargspec(func)

Get the names and default values of a Python function’s arguments. A named tuple ArgSpec(args, varargs, keywords, defaults) is returned. args is a list of the argument names. varargs and keywords are the names of the * and ** arguments or None. defaults is a tuple of default argument values or None if there are no default arguments; if this tuple has n elements, they correspond to the last n elements listed in args.

Deprecated since version 3.0: Use getfullargspec() instead, which provides information about keyword-only arguments and annotations.

inspect.getfullargspec(func)

Get the names and default values of a Python function’s arguments. A named tuple is returned:

FullArgSpec(args, varargs, varkw, defaults, kwonlyargs, kwonlydefaults, annotations)

args is a list of the argument names. varargs and varkw are the names of the * and ** arguments or None. defaults is an n-tuple of the default values of the last n arguments, or None if there are no default arguments. kwonlyargs is a list of keyword-only argument names. kwonlydefaults is a dictionary mapping names from kwonlyargs to defaults. annotations is a dictionary mapping argument names to annotations.

The first four items in the tuple correspond to getargspec().

Note

Consider using the new Signature Object interface, which provides a better way of introspecting functions.

Changed in version 3.4: This function is now based on signature(), but still ignores __wrapped__ attributes and includes the already bound first parameter in the signature output for bound methods.

inspect.getargvalues(frame)

Get information about arguments passed into a particular frame. A named tuple ArgInfo(args, varargs, keywords, locals) is returned. args is a list of the argument names. varargs and keywords are the names of the * and ** arguments or None. locals is the locals dictionary of the given frame.

inspect.formatargspec(args[, varargs, varkw, defaults, kwonlyargs, kwonlydefaults, annotations[, formatarg, formatvarargs, formatvarkw, formatvalue, formatreturns, formatannotations]])

Format a pretty argument spec from the values returned by getargspec() or getfullargspec().

The first seven arguments are (args, varargs, varkw, defaults, kwonlyargs, kwonlydefaults, annotations).

The other six arguments are functions that are called to turn argument names, * argument name, ** argument name, default values, return annotation and individual annotations into strings, respectively.

For example:

>>> from inspect import formatargspec, getfullargspec
>>> def f(a: int, b: float):
...     pass
...
>>> formatargspec(*getfullargspec(f))
'(a: int, b: float)'
inspect.formatargvalues(args[, varargs, varkw, locals, formatarg, formatvarargs, formatvarkw, formatvalue])

Format a pretty argument spec from the four values returned by getargvalues(). The format* arguments are the corresponding optional formatting functions that are called to turn names and values into strings.

inspect.getmro(cls)

Return a tuple of class cls’s base classes, including cls, in method resolution order. No class appears more than once in this tuple. Note that the method resolution order depends on cls’s type. Unless a very peculiar user-defined metatype is in use, cls will be the first element of the tuple.

inspect.getcallargs(func, *args, **kwds)

Bind the args and kwds to the argument names of the Python function or method func, as if it was called with them. For bound methods, bind also the first argument (typically named self) to the associated instance. A dict is returned, mapping the argument names (including the names of the * and ** arguments, if any) to their values from args and kwds. In case of invoking func incorrectly, i.e. whenever func(*args, **kwds) would raise an exception because of incompatible signature, an exception of the same type and the same or similar message is raised. For example:

>>> from inspect import getcallargs
>>> def f(a, b=1, *pos, **named):
...     pass
>>> getcallargs(f, 1, 2, 3) == {'a': 1, 'named': {}, 'b': 2, 'pos': (3,)}
True
>>> getcallargs(f, a=2, x=4) == {'a': 2, 'named': {'x': 4}, 'b': 1, 'pos': ()}
True
>>> getcallargs(f)
Traceback (most recent call last):
...
TypeError: f() missing 1 required positional argument: 'a'

New in version 3.2.

Note

Consider using the new Signature.bind() instead.

inspect.getclosurevars(func)

Get the mapping of external name references in a Python function or method func to their current values. A named tuple ClosureVars(nonlocals, globals, builtins, unbound) is returned. nonlocals maps referenced names to lexical closure variables, globals to the function’s module globals and builtins to the builtins visible from the function body. unbound is the set of names referenced in the function that could not be resolved at all given the current module globals and builtins.

TypeError is raised if func is not a Python function or method.

New in version 3.3.

inspect.unwrap(func, *, stop=None)

Get the object wrapped by func. It follows the chain of __wrapped__ attributes returning the last object in the chain.

stop is an optional callback accepting an object in the wrapper chain as its sole argument that allows the unwrapping to be terminated early if the callback returns a true value. If the callback never returns a true value, the last object in the chain is returned as usual. For example, signature() uses this to stop unwrapping if any object in the chain has a __signature__ attribute defined.

ValueError is raised if a cycle is encountered.

New in version 3.4.

29.12.5. The interpreter stack

When the following functions return “frame records,” each record is a tuple of six items: the frame object, the filename, the line number of the current line, the function name, a list of lines of context from the source code, and the index of the current line within that list.

Note

Keeping references to frame objects, as found in the first element of the frame records these functions return, can cause your program to create reference cycles. Once a reference cycle has been created, the lifespan of all objects which can be accessed from the objects which form the cycle can become much longer even if Python’s optional cycle detector is enabled. If such cycles must be created, it is important to ensure they are explicitly broken to avoid the delayed destruction of objects and increased memory consumption which occurs.

Though the cycle detector will catch these, destruction of the frames (and local variables) can be made deterministic by removing the cycle in a finally clause. This is also important if the cycle detector was disabled when Python was compiled or using gc.disable(). For example:

def handle_stackframe_without_leak():
    frame = inspect.currentframe()
    try:
        # do something with the frame
    finally:
        del frame

If you want to keep the frame around (for example to print a traceback later), you can also break reference cycles by using the frame.clear() method.

The optional context argument supported by most of these functions specifies the number of lines of context to return, which are centered around the current line.

inspect.getframeinfo(frame, context=1)

Get information about a frame or traceback object. A named tuple Traceback(filename, lineno, function, code_context, index) is returned.

inspect.getouterframes(frame, context=1)

Get a list of frame records for a frame and all outer frames. These frames represent the calls that lead to the creation of frame. The first entry in the returned list represents frame; the last entry represents the outermost call on frame‘s stack.

inspect.getinnerframes(traceback, context=1)

Get a list of frame records for a traceback’s frame and all inner frames. These frames represent calls made as a consequence of frame. The first entry in the list represents traceback; the last entry represents where the exception was raised.

inspect.currentframe()

Return the frame object for the caller’s stack frame.

CPython implementation detail: This function relies on Python stack frame support in the interpreter, which isn’t guaranteed to exist in all implementations of Python. If running in an implementation without Python stack frame support this function returns None.

inspect.stack(context=1)

Return a list of frame records for the caller’s stack. The first entry in the returned list represents the caller; the last entry represents the outermost call on the stack.

inspect.trace(context=1)

Return a list of frame records for the stack between the current frame and the frame in which an exception currently being handled was raised in. The first entry in the list represents the caller; the last entry represents where the exception was raised.

29.12.6. Fetching attributes statically

Both getattr() and hasattr() can trigger code execution when fetching or checking for the existence of attributes. Descriptors, like properties, will be invoked and __getattr__() and __getattribute__() may be called.

For cases where you want passive introspection, like documentation tools, this can be inconvenient. getattr_static() has the same signature as getattr() but avoids executing code when it fetches attributes.

inspect.getattr_static(obj, attr, default=None)

Retrieve attributes without triggering dynamic lookup via the descriptor protocol, __getattr__() or __getattribute__().

Note: this function may not be able to retrieve all attributes that getattr can fetch (like dynamically created attributes) and may find attributes that getattr can’t (like descriptors that raise AttributeError). It can also return descriptors objects instead of instance members.

If the instance __dict__ is shadowed by another member (for example a property) then this function will be unable to find instance members.

New in version 3.2.

getattr_static() does not resolve descriptors, for example slot descriptors or getset descriptors on objects implemented in C. The descriptor object is returned instead of the underlying attribute.

You can handle these with code like the following. Note that for arbitrary getset descriptors invoking these may trigger code execution:

# example code for resolving the builtin descriptor types
class _foo:
    __slots__ = ['foo']

slot_descriptor = type(_foo.foo)
getset_descriptor = type(type(open(__file__)).name)
wrapper_descriptor = type(str.__dict__['__add__'])
descriptor_types = (slot_descriptor, getset_descriptor, wrapper_descriptor)

result = getattr_static(some_object, 'foo')
if type(result) in descriptor_types:
    try:
        result = result.__get__()
    except AttributeError:
        # descriptors can raise AttributeError to
        # indicate there is no underlying value
        # in which case the descriptor itself will
        # have to do
        pass

29.12.7. Current State of a Generator

When implementing coroutine schedulers and for other advanced uses of generators, it is useful to determine whether a generator is currently executing, is waiting to start or resume or execution, or has already terminated. getgeneratorstate() allows the current state of a generator to be determined easily.

inspect.getgeneratorstate(generator)

Get current state of a generator-iterator.

Possible states are:
  • GEN_CREATED: Waiting to start execution.
  • GEN_RUNNING: Currently being executed by the interpreter.
  • GEN_SUSPENDED: Currently suspended at a yield expression.
  • GEN_CLOSED: Execution has completed.

New in version 3.2.

The current internal state of the generator can also be queried. This is mostly useful for testing purposes, to ensure that internal state is being updated as expected:

inspect.getgeneratorlocals(generator)

Get the mapping of live local variables in generator to their current values. A dictionary is returned that maps from variable names to values. This is the equivalent of calling locals() in the body of the generator, and all the same caveats apply.

If generator is a generator with no currently associated frame, then an empty dictionary is returned. TypeError is raised if generator is not a Python generator object.

CPython implementation detail: This function relies on the generator exposing a Python stack frame for introspection, which isn’t guaranteed to be the case in all implementations of Python. In such cases, this function will always return an empty dictionary.

New in version 3.3.

29.12.8. Command Line Interface

The inspect module also provides a basic introspection capability from the command line.

By default, accepts the name of a module and prints the source of that module. A class or function within the module can be printed instead by appended a colon and the qualified name of the target object.

--details

Print information about the specified object rather than the source code