Added in version 3.8.
Changed in version 3.10: importlib.metadata is no longer provisional.
Source code:
Lib/importlib/metadata/__init__.py
importlib.metadata is a library that provides access to the metadata of an installed
, such as its entry points or its top-level names (
s, modules, if any). Built in part on Python’s import system, this library provides the entry point and metadata APIs that were previously exposed by the now-removed pkg_resources package. Along with
, it supersedes pkg_resources.
importlib.metadata operates on third-party distribution packages installed into Python’s site-packages directory via tools such as
. Specifically, it works with distributions with discoverable dist-info or egg-info directories, and metadata defined by the
.
Important
These are not necessarily equivalent to or correspond 1:1 with the top-level import package names that can be imported inside Python code. One distribution package can contain multiple import packages (and single modules), and one top-level import package may map to multiple distribution packages if it is a namespace package. You can use
to get a mapping between them.
By default, distribution metadata can live on the file system or in zip archives on
. Through an extension mechanism, the metadata can live almost anywhere.
Overview
Let’s say you wanted to get the version string for a
you’ve installed using pip. We start by creating a virtual environment and installing something into it:
$ python-mvenvexample $ sourceexample/bin/activate (example)$ python-mpipinstallwheel You can get the version string for wheel by running the following:
(example) $ python>>> fromimportlib.metadataimportversion>>> version('wheel')'0.32.3'You can also get a collection of entry points selectable by properties of the EntryPoint (typically ‘group’ or ‘name’), such as console_scripts, distutils.commands and others. Each group contains a collection of
objects.
You can get the
:
>>> list(metadata('wheel'))['Metadata-Version', 'Name', 'Version', 'Summary', 'Home-page', 'Author', 'Author-email', 'Maintainer', 'Maintainer-email', 'License', 'Project-URL', 'Project-URL', 'Project-URL', 'Keywords', 'Platform', 'Classifier', 'Classifier', 'Classifier', 'Classifier', 'Classifier', 'Classifier', 'Classifier', 'Classifier', 'Classifier', 'Classifier', 'Classifier', 'Classifier', 'Requires-Python', 'Provides-Extra', 'Requires-Dist', 'Requires-Dist']You can also get a
, list its
, and get a list of the distribution’s
.
exceptionimportlib.metadata.PackageNotFoundError
Subclass of
raised by several functions in this module when queried for a distribution package which is not installed in the current Python environment.
exceptionimportlib.metadata.MetadataNotFound
Subclass of
raised when attempting to load metadata from a distribution folder that is empty or otherwise does not contain a metadata file.
Functional API
This package provides the following functionality via its public API.
Entry points
importlib.metadata.entry_points(**select_params)
Returns a
instance describing entry points for the current environment. Any given keyword parameters are passed to the select() method for comparison to the attributes of the individual entry point definitions.
Note: to query for entry points based on EntryPoint.dist attribute, use
instead (as different
instances do not currently compare equal, even if they have the same attributes)
classimportlib.metadata.EntryPoints
Details of a collection of installed entry points.
Also provides a .groups attribute that reports all identified entry point groups, and a .names attribute that reports all identified entry point names.
classimportlib.metadata.EntryPoint
Details of an installed entry point.
Each EntryPoint instance has .name, .group, and .value attributes and a .load() method to resolve the value. There are also .module, .attr, and .extras attributes for getting the components of the .value attribute, and .dist for obtaining information regarding the distribution package that provides the entry point.
Query all entry points:
>>> eps=entry_points()The entry_points() function returns a EntryPoints object, a collection of all EntryPoint objects with names and groups attributes for convenience:
>>> sorted(eps.groups)['console_scripts', 'distutils.commands', 'distutils.setup_keywords', 'egg_info.writers', 'setuptools.installation']EntryPoints has a select() method to select entry points matching specific properties. Select entry points in the console_scripts group:
>>> scripts=eps.select(group='console_scripts')Equivalently, since entry_points() passes keyword arguments through to select:
>>> scripts=entry_points(group='console_scripts')Pick out a specific script named “wheel” (found in the wheel project):
>>> 'wheel'inscripts.namesTrue>>> wheel=scripts['wheel']Equivalently, query for that entry point during selection:
>>> (wheel,)=entry_points(group='console_scripts',name='wheel')>>> (wheel,)=entry_points().select(group='console_scripts',name='wheel')Inspect the resolved entry point:
>>> wheelEntryPoint(name='wheel', value='wheel.cli:main', group='console_scripts')>>> wheel.module'wheel.cli'>>> wheel.attr'main'>>> wheel.extras[]>>> main=wheel.load()>>> main<function main at 0x103528488>The group and name are arbitrary values defined by the package author and usually a client will wish to resolve all entry points for a particular group. Read
for more information on entry points, their definition, and usage.
Changed in version 3.12: The “selectable” entry points were introduced in importlib_metadata 3.6 and Python 3.10. Prior to those changes, entry_points accepted no parameters and always returned a dictionary of entry points, keyed by group. With importlib_metadata 5.0 and Python 3.12, entry_points always returns an EntryPoints object. See
backports.entry_points_selectable
for compatibility options.
Changed in version 3.13: EntryPoint objects no longer present a tuple-like interface (
).
Distribution metadata
importlib.metadata.metadata(distribution_name)
Return the distribution metadata corresponding to the named distribution package as a
instance.
Raises
if the named distribution package is not installed in the current Python environment.
Raises
if a distribution package is present but no METADATA file is present.
classimportlib.metadata.PackageMetadata
A concrete implementation of the
.
In addition to providing the defined protocol methods and attributes, subscripting the instance is equivalent to calling the get() method.
Every
includes some metadata, which you can extract using the metadata() function:
>>> wheel_metadata=metadata('wheel')The keys of the returned data structure name the metadata keywords, and the values are returned unparsed from the distribution metadata:
>>> wheel_metadata['Requires-Python']'>=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*'
also presents a json attribute that returns all the metadata in a JSON-compatible form per
:
>>> wheel_metadata.json['requires_python']'>=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*'The full set of available metadata is not described here. See the PyPA
for additional details.
Changed in version 3.15: Previously and incidentally, if a METADATA file was missing from a distribution, an empty PackageMetadata would be returned, indistinguishable from an empty METADATA file. Now, a missing METADATA file triggers a MetadataNotFound exception.
Changed in version 3.10: The Description is now included in the metadata when presented through the payload. Line continuation characters have been removed.
The json attribute was added.
Distribution versions
importlib.metadata.version(distribution_name)
Return the installed distribution package
for the named distribution package.
Raises
if the named distribution package is not installed in the current Python environment.
The version() function is the quickest way to get a
’s version number, as a string:
>>> version('wheel')'0.32.3'Distribution files
importlib.metadata.files(distribution_name)
Return the full set of files contained within the named distribution package as
instances.
Raises
if the named distribution package is not installed in the current Python environment.
Returns
if the distribution is found but the installation database records reporting the files associated with the distribution package are missing.
classimportlib.metadata.PackagePath
A
derived object with additional dist, size, and hash properties corresponding to the distribution package’s installation metadata for that file, also:
locate()
If possible, return the concrete
allowing to access data, or raise a
otherwise.
classimportlib.metadata.SimplePath
A protocol representing a minimal subset of
that allows to check if it exists(), to traverse using joinpath() and parent, and to retrieve data using read_text() and read_bytes().
The files() function takes a
name and returns all of the files installed by this distribution. For example:
>>> util=[pforpinfiles('wheel')if'util.py'instr(p)][0]>>> utilPackagePath('wheel/util.py')>>> util.size859>>> util.dist<importlib.metadata._hooks.PathDistribution object at 0x101e0cef0>>>> util.hash<FileHash mode: sha256 value: bYkw5oMccfazVCoYQwKkkemoVyMAFoR34mmKBx8R1NI>Once you have the file, you can also read its contents:
>>> print(util.read_text())import base64import sys...def as_bytes(s): if isinstance(s, text_type): return s.encode('utf-8') return sYou can also use the locate() method to get the absolute path to the file:
>>> util.locate()PosixPath('/home/gustav/example/lib/site-packages/wheel/util.py')In the case where the metadata file listing files (RECORD or SOURCES.txt) is missing, files() will return
. The caller may wish to wrap calls to files() in
or otherwise guard against this condition if the target distribution is not known to have the metadata present.
Distribution requirements
importlib.metadata.requires(distribution_name)
Return the declared dependency specifiers for the named distribution package.
Raises
if the named distribution package is not installed in the current Python environment.
To get the full set of requirements for a
, use the requires() function:
>>> requires('wheel')["pytest (>=3.0.0) ; extra == 'test'", "pytest-cov ; extra == 'test'"]Mapping import to distribution packages
importlib.metadata.packages_distributions()
Return a mapping from the top level module and import package names found via
to the names of the distribution packages (if any) that provide the corresponding files.
To allow for namespace packages (which may have members provided by multiple distribution packages), each top level import name maps to a list of distribution names rather than mapping directly to a single name.
A convenience method to resolve the
name (or names, in the case of a namespace package) that provide each importable top-level Python module or
:
>>> packages_distributions(){'importlib_metadata': ['importlib-metadata'], 'yaml': ['PyYAML'], 'jaraco': ['jaraco.classes', 'jaraco.functools'], ...}Some editable installs,
, and thus this function is not reliable with such installs.
Added in version 3.10.
Distributions
While the module level API described above is the most common and convenient usage, all that information is accessible from the
class. Distribution is an abstract object that represents the metadata for a Python
. Get the concrete Distribution subclass instance for an installed distribution package by calling the
function:
>>> fromimportlib.metadataimportdistribution>>> dist=distribution('wheel')>>> type(dist)<class 'importlib.metadata.PathDistribution'>importlib.metadata.distribution(distribution_name)
Return a
instance describing the named distribution package.
Raises
if the named distribution package is not installed in the current Python environment.
Thus, an alternative way to get e.g. the version number is through the
attribute:
>>> dist.version'0.32.3'The same applies for
and
.
classimportlib.metadata.Distribution
Details of an installed distribution package.
Note: different Distribution instances do not currently compare equal, even if they relate to the same installed distribution and accordingly have the same attributes.
staticat(path)
classmethodfrom_name(name)
Return a Distribution instance at the given path or with the given name.
classmethoddiscover(*, context=None, **kwargs)
Returns an iterable of Distribution instances for all packages (see
).
The optional argument context is a
instance, used to modify the search for distributions. Alternatively, kwargs may contain keyword arguments for constructing a new DistributionFinder.Context.
metadata:
Raises
if the METADATA file is not present in the distribution.
There are all kinds of additional metadata available on Distribution instances as a
instance:
>>> dist.metadata['Requires-Python']'>=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*'>>> dist.metadata['License']'MIT'The full set of available metadata is not described here. See the PyPA
for additional details.
name:
requires:
[
]
version:
A few metadata fields are also available as shortcut properties.
Added in version 3.10: The name shortcut was added.
origin
For editable packages, an origin property may present
metadata (for non-editable packages, origin is
):
>>> dist.origin.url'file:///path/to/wheel-0.32.3.editable-py3-none-any.whl'The origin object follows the
.
Added in version 3.13.
entry_points:
The entry points provided by this distribution package.
files:
[
]|
All files contained in this distribution package. Like
, this returns
if there are no records.
The following two abstract methods need to be implemented when
:
locate_file(path)
Like PackagePath.locate(), return a
for the given path. Takes a
or a
.
read_text(filename)
A shortcut for distribution.locate_file(filename).read_text().
Distribution Discovery
By default, this package provides built-in support for discovery of metadata for file system and zip file
s. This metadata finder search defaults to sys.path, but varies slightly in how it interprets those values from how other import machinery does. In particular:
importlib.metadata does not honor
objects on sys.path.
importlib.metadata will incidentally honor
objects on sys.path even though such values will be ignored for imports.
classimportlib.metadata.DistributionFinder
A
subclass capable of discovering installed distributions.
Custom providers should implement this interface in order to supply metadata.
classContext(**kwargs)
A Context gives a custom provider a means to solicit additional details from the callers of distribution discovery functions like
or
beyond .name and .path when searching for distributions.
For example, a provider could expose suites of packages in either a “public” or “private” realm. A caller of distribution discovery functions may wish to query only for distributions in a particular realm and could call distributions(realm="private") to signal to the custom provider to only include distributions from that realm.
Each DistributionFinder must expect any parameters and should attempt to honor the canonical parameters defined below when appropriate.
See the section on
for more details.
name
Specific name for which a distribution finder should match.
A .name of None matches all distributions.
path
A property providing the sequence of directory paths that a distribution finder should search.
Typically refers to Python installed package paths such as “site-packages” directories and defaults to
.
importlib.metadata.distributions(**kwargs)
Returns an iterable of
instances for all packages.
The kwargs argument may contain either a keyword argument context, a
instance, or pass keyword arguments for constructing a new DistributionFinder.Context. The DistributionFinder.Context is used to modify the search for distributions.
Implementing Custom Providers
importlib.metadata address two API surfaces, one for consumers and another for providers. Most users are consumers, consuming metadata provided by the packages. There are other use-cases, however, where users wish to expose metadata through some other mechanism, such as alongside a custom importer. Such a use case calls for a custom provider.
Because
metadata is not available through
searches, or package loaders directly, the metadata for a distribution is found through import system
. To find a distribution package’s metadata, importlib.metadata queries the list of
on
.
The implementation has hooks integrated into the PathFinder, serving metadata for distribution packages found on the file system.
The abstract class
defines the interface expected of finders by Python’s import system. importlib.metadata extends this protocol by looking for an optional find_distributions callable on the finders from
and presents this extended interface as the
abstract base class, which defines this abstract method:
@abc.abstractmethoddeffind_distributions(context=DistributionFinder.Context())->Iterable[Distribution]:"""Return an iterable of all Distribution instances capable of loading the metadata for packages for the indicated ``context``. """The
object provides
and
properties indicating the path to search and name to match and may supply other relevant context sought by the consumer.
In practice, to support finding distribution package metadata in locations other than the file system, subclass Distribution and implement the abstract methods. Then from a custom finder, return instances of this derived Distribution in the find_distributions() method.
Example
Imagine a custom finder that loads Python modules from a database:
classDatabaseImporter(importlib.abc.MetaPathFinder):def__init__(self,db):self.db=dbdeffind_spec(self,fullname,target=None)->ModuleSpec:returnself.db.spec_from_name(fullname)sys.meta_path.append(DatabaseImporter(connect_db(...)))That importer now presumably provides importable modules from a database, but it provides no metadata or entry points. For this custom importer to provide metadata, it would also need to implement
:
fromimportlib.metadataimportDistributionFinderclassDatabaseImporter(DistributionFinder):...deffind_distributions(self,context=DistributionFinder.Context()):query=dict(name=context.name)ifcontext.nameelse{}fordist_recordinself.db.query_distributions(query):yieldDatabaseDistribution(dist_record)In this way, query_distributions would return records for each distribution served by the database matching the query. For example, if requests-1.0 is in the database, find_distributions would yield a DatabaseDistribution for Context(name='requests') or Context(name=None).
For the sake of simplicity, this example ignores context.path. The path attribute defaults to sys.path and is the set of import paths to be considered in the search. A DatabaseImporter could potentially function without any concern for a search path. Assuming the importer does no partitioning, the “path” would be irrelevant. In order to illustrate the purpose of path, the example would need to illustrate a more complex DatabaseImporter whose behavior varied depending on sys.path/PYTHONPATH. In that case, the find_distributions should honor the context.path and only yield Distributions pertinent to that path.
DatabaseDistribution, then, would look something like:
classDatabaseDistribution(importlib.metadata.Distribution):def__init__(self,record):self.record=recorddefread_text(self,filename):""" Read a file like "METADATA" for the current distribution. """iffilename=="METADATA":returnf"""Name: {self.record.name}Version: {self.record.version}"""iffilename=="entry_points.txt":return"\n".join(f"""[{ep.group}]\n{ep.name}={ep.value}"""forepinself.record.entry_points)deflocate_file(self,path):raiseRuntimeError("This distribution has no file system")This basic implementation should provide metadata and entry points for packages served by the DatabaseImporter, assuming that the record supplies suitable .name, .version, and .entry_points attributes.
The DatabaseDistribution may also provide other metadata files, like RECORD (required for Distribution.files) or override the implementation of Distribution.files. See the source for more inspiration.