Data Files Support#
Old packaging installation methods in the Python ecosystem have traditionally allowed installation of “data files”, which are placed in a platform-specific location. However, the most common use case for data files distributed with a package is for use by the package, usually by including the data files inside the package directory.
Setuptools focuses on this most common type of data files and offers three ways of specifying which files should be included in your packages, as described in the following sections.
include_package_data#
First, you can simply use the include_package_data keyword.
For example, if the package tree looks like this:
project_root_directory
├── setup.py # and/or setup.cfg, pyproject.toml
└── src
└── mypkg
├── __init__.py
├── data1.rst
├── data2.rst
├── data1.txt
└── data2.txt
and you supply this configuration:
[tool.setuptools]
# ...
# By default, include-package-data is true in pyproject.toml, so you do
# NOT have to specify this line.
include-package-data = true
[tool.setuptools.packages.find]
where = ["src"]
[options]
# ...
packages = find:
package_dir =
= src
include_package_data = True
[options.packages.find]
where = src
from setuptools import setup, find_packages
setup(
# ...,
packages=find_packages(where="src"),
package_dir={"": "src"},
include_package_data=True
)
then all the .txt and .rst files will be automatically installed with
your package, provided:
These files are included via the MANIFEST.in file, like so:
include src/mypkg/*.txt include src/mypkg/*.rst
OR, they are being tracked by a revision control system such as Git, Mercurial or SVN, and you have configured an appropriate plugin such as setuptools-scm or setuptools-svn. (See the section below on Adding Support for Revision Control Systems for information on how to write such plugins.)
Note
New in version v61.0.0: The default value for tool.setuptools.include-package-data is True
when projects are configured via pyproject.toml.
This behaviour differs from setup.cfg and setup.py
(where include_package_data=False by default), which was not changed
to ensure backwards compatibility with existing projects.
package_data#
By default, include_package_data considers all non .py files found inside
the package directory (src/mypkg in this case) as data files, and includes those that
satisfy (at least) one of the above two conditions into the source distribution, and
consequently in the installation of your package.
If you want finer-grained control over what files are included, then you can also use
the package_data keyword.
For example, if the package tree looks like this:
project_root_directory
├── setup.py # and/or setup.cfg, pyproject.toml
└── src
└── mypkg
├── __init__.py
├── data1.rst
├── data2.rst
├── data1.txt
└── data2.txt
then you can use the following configuration to capture the .txt and .rst files as
data files:
[tool.setuptools.packages.find]
where = ["src"]
[tool.setuptools.package-data]
mypkg = ["*.txt", "*.rst"]
[options]
# ...
packages = find:
package_dir =
= src
[options.packages.find]
where = src
[options.package_data]
mypkg =
*.txt
*.rst
from setuptools import setup, find_packages
setup(
# ...,
packages=find_packages(where="src"),
package_dir={"": "src"},
package_data={"mypkg": ["*.txt", "*.rst"]}
)
The package_data argument is a dictionary that maps from package names to
lists of glob patterns. Note that the data files specified using the package_data
option neither require to be included within a MANIFEST.in
file, nor require to be added by a revision control system plugin.
Note
If your glob patterns use paths, you must use a forward slash (/) as
the path separator, even if you are on Windows. Setuptools automatically
converts slashes to appropriate platform-specific separators at build time.
Important
Glob patterns do not automatically match dotfiles, i.e., directory or file names
starting with a dot (.). To include such files, you must explicitly start
the pattern with a dot, e.g. .* to match .gitignore.
If you have multiple top-level packages and a common pattern of data files for all these packages, for example:
project_root_directory
├── setup.py # and/or setup.cfg, pyproject.toml
└── src
├── mypkg1
│ ├── data1.rst
│ ├── data1.txt
│ └── __init__.py
└── mypkg2
├── data2.txt
└── __init__.py
Here, both packages mypkg1 and mypkg2 share a common pattern of having .txt
data files. However, only mypkg1 has .rst data files. In such a case, if you want to
use the package_data option, the following configuration will work:
[tool.setuptools.packages.find]
where = ["src"]
[tool.setuptools.package-data]
"*" = ["*.txt"]
mypkg1 = ["data1.rst"]
[options]
packages = find:
package_dir =
= src
[options.packages.find]
where = src
[options.package_data]
* =
*.txt
mypkg1 =
data1.rst
from setuptools import setup, find_packages
setup(
# ...,
packages=find_packages(where="src"),
package_dir={"": "src"},
package_data={"": ["*.txt"], "mypkg1": ["data1.rst"]},
)
Notice that if you list patterns in package_data under the empty string "" in
setup.py, and the asterisk * in setup.cfg and pyproject.toml, these
patterns are used to find files in every package. For example, we use "" or *
to indicate that the .txt files from all packages should be captured as data files.
These placeholders are treated as a special case, setuptools do not
support glob patterns on package names for this configuration
(patterns are only supported on the file paths).
Also note how we can continue to specify patterns for individual packages, i.e.
we specify that data1.rst from mypkg1 alone should be captured as well.
Note
When building an sdist, the data files are also drawn from the
package_name.egg-info/SOURCES.txt file which works as a form of cache.
So make sure that this file is removed if package_data is updated,
before re-building the package.
Attention
In Python any directory is considered a package
(even if it does not contain __init__.py,
see native namespaces packages on Packaging namespace packages).
Therefore, if you are not relying on automatic discovery,
you SHOULD ensure that all packages (including the ones that don’t
contain any Python files) are included in the packages configuration
(see Package Discovery and Namespace Packages for more information).
Moreover, it is advisable to use full packages name using the dot notation instead of a nested path, to avoid error prone configurations. Please check section subdirectories below.
exclude_package_data#
Sometimes, the include_package_data or package_data options alone
aren’t sufficient to precisely define what files you want included. For example,
consider a scenario where you have include_package_data=True, and you are using
a revision control system with an appropriate plugin.
Sometimes developers add directory-specific marker files (such as .gitignore,
.gitkeep, .gitattributes, or .hgignore), these files are probably being
tracked by the revision control system, and therefore by default they will be
included when the package is installed.
Supposing you want to prevent these files from being included in the
installation (they are not relevant to Python or the package), then you could
use the exclude_package_data option:
[tool.setuptools.packages.find]
where = ["src"]
[tool.setuptools.exclude-package-data]
mypkg = [".gitattributes"]
[options]
# ...
packages = find:
package_dir =
= src
include_package_data = True
[options.packages.find]
where = src
[options.exclude_package_data]
mypkg =
.gitattributes
from setuptools import setup, find_packages
setup(
# ...,
packages=find_packages(where="src"),
package_dir={"": "src"},
include_package_data=True,
exclude_package_data={"mypkg": [".gitattributes"]},
)
The exclude_package_data option is a dictionary mapping package names to
lists of wildcard patterns, just like the package_data option. And, just
as with that option, you can use the empty string key "" in setup.py and the
asterisk * in setup.cfg and pyproject.toml to match all top-level packages.
Any files that match these patterns will be excluded from installation,
even if they were listed in package_data or were included as a result of using
include_package_data.
Subdirectory for Data Files#
A common pattern is where some (or all) of the data files are placed under a separate subdirectory. For example:
project_root_directory
├── setup.py # and/or setup.cfg, pyproject.toml
└── src
└── mypkg
├── data
│ ├── data1.rst
│ └── data2.rst
├── __init__.py
├── data1.txt
└── data2.txt
Here, the .rst files are placed under a data subdirectory inside mypkg,
while the .txt files are directly under mypkg.
In this case, the recommended approach is to treat data as a namespace package
(refer PEP 420). With package_data,
the configuration might look like this:
# Scanning for namespace packages in the ``src`` directory is true by
# default in pyproject.toml, so you do NOT need to include the
# `tool.setuptools.packages.find` if it looks like the following:
# [tool.setuptools.packages.find]
# namespaces = true
# where = ["src"]
[tool.setuptools.package-data]
mypkg = ["*.txt"]
"mypkg.data" = ["*.rst"]
[options]
# ...
packages = find_namespace:
package_dir =
= src
[options.packages.find]
where = src
[options.package_data]
mypkg =
*.txt
mypkg.data =
*.rst
from setuptools import setup, find_namespace_packages
setup(
# ...,
packages=find_namespace_packages(where="src"),
package_dir={"": "src"},
package_data={
"mypkg": ["*.txt"],
"mypkg.data": ["*.rst"],
}
)
In other words, we allow Setuptools to scan for namespace packages in the src directory,
which enables the data directory to be identified, and then, we separately specify data
files for the root package mypkg, and the namespace package data under the package
mypkg.
With include_package_data the configuration is simpler: you simply need to enable
scanning of namespace packages in the src directory and the rest is handled by Setuptools.
[tool.setuptools]
# ...
# By default, include-package-data is true in pyproject.toml, so you do
# NOT have to specify this line.
include-package-data = true
[tool.setuptools.packages.find]
# scanning for namespace packages is true by default in pyproject.toml, so
# you need NOT include the following line.
namespaces = true
where = ["src"]
[options]
packages = find_namespace:
package_dir =
= src
include_package_data = True
[options.packages.find]
where = src
from setuptools import setup, find_namespace_packages
setup(
# ... ,
packages=find_namespace_packages(where="src"),
package_dir={"": "src"},
include_package_data=True,
)
Summary#
In summary, the three options allow you to:
include_package_dataAccept all data files and directories matched by MANIFEST.in or added by a plugin.
package_dataSpecify additional patterns to match files that may or may not be matched by MANIFEST.in or added by a plugin.
exclude_package_dataSpecify patterns for data files and directories that should not be included when a package is installed, even if they would otherwise have been included due to the use of the preceding options.
Note
Due to the way the build process works, a data file that you include in your project and then stop including may be “orphaned” in your project’s build directories, requiring you to manually deleting them. This may also be important for your users and contributors if they track intermediate revisions of your project using Subversion; be sure to let them know when you make changes that remove files from inclusion so they can also manually delete them.
Accessing Data Files at Runtime#
Typically, existing programs manipulate a package’s __file__ attribute in
order to find the location of data files. For example, if you have a structure
like this:
project_root_directory
├── setup.py # and/or setup.cfg, pyproject.toml
└── src
└── mypkg
├── data
│ └── data1.txt
├── __init__.py
└── foo.py
Then, in mypkg/foo.py, you may try something like this in order to access
mypkg/data/data1.txt:
import os
data_path = os.path.join(os.path.dirname(__file__), 'data', 'data1.txt')
with open(data_path, 'r') as data_file:
...
However, this manipulation isn’t compatible with PEP 302-based import hooks,
including importing from zip files and Python Eggs. It is strongly recommended that,
if you are using data files, you should use importlib.resources to access them.
In this case, you would do something like this:
from importlib.resources import files
data_text = files('mypkg.data').joinpath('data1.txt').read_text()
importlib.resources was added to Python 3.7. However, the API illustrated in
this code (using files()) was added only in Python 3.9, [2] and support
for accessing data files via namespace packages was added only in Python 3.10 [3]
(the data subdirectory is a namespace package under the root package mypkg).
Therefore, you may find this code to work only in Python 3.10 (and above). For other
versions of Python, you are recommended to use the importlib-resources backport
which provides the latest version of this library. In this case, the only change that
has to be made to the above code is to replace importlib.resources with importlib_resources, i.e.
from importlib_resources import files
...
See Using importlib_resources for detailed instructions.
Tip
Files inside the package directory should be read-only to avoid a series of common problems (e.g. when multiple users share a common Python installation, when the package is loaded from a zip file, or when multiple instances of a Python application run in parallel).
If your Python package needs to write to a file for shared data or configuration,
you can use standard platform/OS-specific system directories, such as
~/.local/config/$appname or /usr/share/$appname/$version (Linux specific) [1].
A common approach is to add a read-only template file to the package
directory that is then copied to the correct system directory if no
pre-existing file is found.
Data Files from Plugins and Extensions#
You can resort to a native/implicit namespace package (as a container for files)
if you want plugins and extensions to your package to contribute with package data files.
This way, all files will be listed during runtime
when using importlib.resources.
Note that, although not strictly guaranteed, mainstream Python package managers,
like pip and derived tools, will install files belong to multiple distributions
that share a same namespace into the same directory in the file system.
This means that the overhead for importlib.resources will be minimum.
Non-Package Data Files#
Historically, setuptools by way of easy_install would encapsulate data
files from the distribution into the egg (see the old docs). As eggs are deprecated and pip-based installs
fall back to the platform-specific location for installing data files, there is
no supported facility to reliably retrieve these resources.
Instead, the PyPA recommends that any data files you wish to be accessible at run time be included inside the package.