The pyproject.toml file

The `pyproject.toml` file #

The tool.poetry section of the pyproject.toml file is composed of multiple sections.

package-mode #

Whether Poetry operates in package mode (default) or not. Optional

See basic usage for more information.

package-mode = false

name #

The name of the package. Required in package mode

This should be a valid name as defined by PEP 508.

name = "my-package"

version #

The version of the package. Required in package mode

This should be a valid PEP 440 string.

version = "0.1.0"

Note

If you would like to use semantic versioning for your project, please see here.

description #

A short description of the package. Required in package mode

description = "A short description of the package."

license #

The license of the package.

The recommended notation for the most common licenses is (alphabetical):

Apache-2.0
BSD-2-Clause
BSD-3-Clause
BSD-4-Clause
GPL-2.0-only
GPL-2.0-or-later
GPL-3.0-only
GPL-3.0-or-later
LGPL-2.1-only
LGPL-2.1-or-later
LGPL-3.0-only
LGPL-3.0-or-later
MIT

Optional, but it is highly recommended to supply this. More identifiers are listed at the SPDX Open Source License Registry.

license = "MIT"

Note

If your project is proprietary and does not use a specific licence, you can set this value as Proprietary.

authors #

The authors of the package. Required in package mode

This is a list of authors and should contain at least one author. Authors must be in the form name <email>.

authors = [
    "Sébastien Eustace <[email protected]>",
]

maintainers #

The maintainers of the package. Optional

This is a list of maintainers and should be distinct from authors. Maintainers may contain an email and be in the form name <email>.

maintainers = [
    "John Smith <[email protected]>",
    "Jane Smith <[email protected]>",
]

readme #

A path, or list of paths corresponding to the README file(s) of the package. Optional

The file(s) can be of any format, but if you intend to publish to PyPI keep the recommendations for a PyPI-friendly README in mind. README paths are implicitly relative to pyproject.toml.

Note

Whether paths are case-sensitive follows platform defaults, but it is recommended to keep cases.

To be specific, you can set readme = "rEaDmE.mD" for README.md on macOS and Windows, but Linux users can’t poetry install after cloning your repo. This is because macOS and Windows are case-insensitive and case-preserving.

The contents of the README file(s) are used to populate the Description field of your distribution’s metadata (similar to long_description in setuptools). When multiple files are specified they are concatenated with newlines.

[tool.poetry]
# ...
readme = "README.md"

[tool.poetry]
# ...
readme = ["docs/README1.md", "docs/README2.md"]

homepage #

An URL to the website of the project. Optional

homepage = "https://python-poetry.org/"

repository #

An URL to the repository of the project. Optional

repository = "https://github.com/python-poetry/poetry"

documentation #

An URL to the documentation of the project. Optional

documentation = "https://python-poetry.org/docs/"

keywords #

A list of keywords that the package is related to. Optional

keywords = ["packaging", "poetry"]

classifiers #

A list of PyPI trove classifiers that describe the project. Optional

[tool.poetry]
# ...
classifiers = [
    "Topic :: Software Development :: Build Tools",
    "Topic :: Software Development :: Libraries :: Python Modules"
]

Note

Note that Python classifiers are still automatically added for you and are determined by your python requirement.

The license property will also set the License classifier automatically.

packages #

A list of packages and modules to include in the final distribution.

If your project structure differs from the standard one supported by poetry, you can specify the packages you want to include in the final distribution.

[tool.poetry]
# ...
packages = [
    { include = "my_package" },
    { include = "extra_package/**/*.py" },
]

If your package is stored inside a “lib” directory, you must specify it:

[tool.poetry]
# ...
packages = [
    { include = "my_package", from = "lib" },
]

The to parameter is designed to specify the relative destination path where the package will be located upon installation. This allows for greater control over the organization of packages within your project’s structure.

[tool.poetry]
# ...
packages = [
    { include = "my_package", from = "lib", to = "target_package" },
]

If you want to restrict a package to a specific build format you can specify it by using format:

[tool.poetry]
# ...
packages = [
    { include = "my_package" },
    { include = "my_other_package", format = "sdist" },
]

From now on, only the sdist build archive will include the my_other_package package.

Note

Using packages disables the package auto-detection feature meaning you have to explicitly specify the “default” package.

For instance, if you have a package named my_package and you want to also include another package named extra_package, you will need to specify my_package explicitly:

packages = [
    { include = "my_package" },
    { include = "extra_package" },
]

Note

Poetry is clever enough to detect Python subpackages.

Thus, you only have to specify the directory where your root package resides.

include and exclude #

A list of patterns that will be included in the final package.

You can explicitly specify to Poetry that a set of globs should be ignored or included for the purposes of packaging. The globs specified in the exclude field identify a set of files that are not included when a package is built.

If a VCS is being used for a package, the exclude field will be seeded with the VCS’ ignore settings (.gitignore for git for example).

Note

Explicitly declaring entries in include will negate VCS' ignore settings.

[tool.poetry]
# ...
include = ["CHANGELOG.md"]

You can also specify the formats for which these patterns have to be included, as shown here:

[tool.poetry]
# ...
include = [
    { path = "tests", format = "sdist" },
    { path = "for_wheel.txt", format = ["sdist", "wheel"] }
]

If no format is specified, include defaults to only sdist.

In contrast, exclude defaults to both sdist and wheel.

exclude = ["my_package/excluded.py"]

dependencies and dependency groups #

Poetry is configured to look for dependencies on PyPI by default. Only the name and a version string are required in this case.

[tool.poetry.dependencies]
requests = "^2.13.0"

If you want to use a private repository, you can add it to your pyproject.toml file, like so:

[[tool.poetry.source]]
name = "private"
url = "https://example.com/simple"

If you have multiple repositories configured, you can explicitly tell poetry where to look for a specific package:

[tool.poetry.dependencies]
requests = { version = "^2.13.0", source = "private" }

Note

Be aware that declaring the python version for which your package is compatible is mandatory:

[tool.poetry.dependencies]
python = "^3.7"

You can organize your dependencies in groups to manage them in a more granular way.

[tool.poetry.group.test.dependencies]
pytest = "*"

[tool.poetry.group.docs.dependencies]
mkdocs = "*"

See Dependency groups for a more in-depth look at how to manage dependency groups and Dependency specification for more information on other keys and specifying version ranges.

`scripts` #

This section describes the scripts or executables that will be installed when installing the package

[tool.poetry.scripts]
my_package_cli = 'my_package.console:run'

Here, we will have the my_package_cli script installed which will execute the run function in the console module in the my_package package.

Note

When a script is added or updated, run poetry install to make them available in the project’s virtualenv.

`extras` #

Poetry supports extras to allow expression of:

optional dependencies, which enhance a package, but are not required; and
clusters of optional dependencies.

[tool.poetry]
name = "awesome"

[tool.poetry.dependencies]
# These packages are mandatory and form the core of this package’s distribution.
mandatory = "^1.0"

# A list of all of the optional dependencies, some of which are included in the
# below `extras`. They can be opted into by apps.
psycopg2 = { version = "^2.9", optional = true }
mysqlclient = { version = "^1.3", optional = true }

[tool.poetry.extras]
mysql = ["mysqlclient"]
pgsql = ["psycopg2"]
databases = ["mysqlclient", "psycopg2"]

When installing packages with Poetry, you can specify extras by using the -E|--extras option:

poetry install --extras "mysql pgsql"
poetry install -E mysql -E pgsql

Any extras you don’t specify will be removed. Note this behavior is different from optional dependency groups not selected for install, e.g. those not specified via install --with.

You can install all extras with the --all-extras option:

poetry install --all-extras

Note

Note that install --extras and the variations mentioned above (--all-extras, --extras foo, etc.) only work on dependencies defined in the current project. If you want to install extras defined by dependencies, you’ll have to express that in the dependency itself:

[tool.poetry.dependencies]
pandas = {version="^2.2.1", extras=["computation", "performance"]}

[tool.poetry.group.dev.dependencies]
fastapi = {version="^0.92.0", extras=["all"]}

When installing or specifying Poetry-built packages, the extras defined in this section can be activated as described in PEP 508.

For example, when installing the package using pip, the dependencies required by the databases extra can be installed as shown below.

pip install awesome[databases]

Note

The dependencies specified for each extra must already be defined as project dependencies.

Dependencies listed in dependency groups cannot be specified as extras.

`plugins` #

Poetry supports arbitrary plugins, which are exposed as the ecosystem-standard entry points and discoverable using importlib.metadata. This is similar to (and compatible with) the entry points feature of setuptools. The syntax for registering a plugin is:

[tool.poetry.plugins] # Optional super table

[tool.poetry.plugins."A"]
B = "C:D"

Which are:

A - type of the plugin, for example poetry.plugin or flake8.extension
B - name of the plugin
C - python module import path
D - the entry point of the plugin (a function or class)

Example (from poetry-plugin-export):

[tool.poetry.plugins."poetry.application.plugin"]
export = "poetry_plugin_export.plugins:ExportApplicationPlugin"

`urls` #

In addition to the basic urls (homepage, repository and documentation), you can specify any custom url in the urls section.

[tool.poetry.urls]
"Bug Tracker" = "https://github.com/python-poetry/poetry/issues"

If you publish your package on PyPI, they will appear in the Project Links section.

Poetry and PEP-517 #

PEP-517 introduces a standard way to define alternative build systems to build a Python project.

Poetry is compliant with PEP-517, by providing a lightweight core library, so if you use Poetry to manage your Python project you should reference it in the build-system section of the pyproject.toml file like so:

[build-system]
requires = ["poetry-core>=1.0.0"]
build-backend = "poetry.core.masonry.api"

Note

When using the new or init command this section will be automatically added.

Note

If your pyproject.toml file still references poetry directly as a build backend, you should update it to reference poetry-core instead.

The pyproject.toml file

The pyproject.toml file #

package-mode #

name #

version #

description #

license #

authors #

maintainers #

readme #

homepage #

repository #

documentation #

keywords #

classifiers #

packages #

include and exclude #

dependencies and dependency groups #

scripts #

extras #

plugins #

urls #

Poetry and PEP-517 #

The `pyproject.toml` file #

`scripts` #

`extras` #

`plugins` #

`urls` #