Skip to content

Packaging & deploying


Packaging and deploying Polylith projects is done by using the Poetry Multiproject plugin command (see installation).

The poetry build-project command will make it possible to use relative package includes as how components and bases are added to Python Polylith projects. Relative includes are currently not possible by default in Poetry, that is where the Multiproject plugin comes in.

Hatch, PDM and Rye

Hatch and PDM support relative includes via the [tool.poetry.bricks] configuration. Nothing extra needed other than the build hooks.

Building source distributions (sdist)?

If you will distribute an sdist as the primary way to install your package, you will need to add the path in the project-specific pyroject.toml.

If you only provide wheel distributions, this is optional.

Hatch and Rye

packages = ["<your top namespace here>"]


includes = ["<your top namespace here>/"]


To collect the components and bases that are needed for a specific project, the tool introduces a build step. The tool will build a wheel and an sdist from the source code of a project.

This is the preferred way for Polylith projects.

Packaging a service or app


poetry build-project --directory path/to/project


cd path/to_project

hatch build


cd path/to_project

pdm build


cd path/to_project

# run the build command for each output
rye build --wheel
rye build --sdist

This command will create a project specific dist folder containing a wheel and an sdist. You can use the available build options with this command too.


You can use the built artifacts to install your service in your preffered way, just by running

pip install the-built-artifact.whl

Packaging a Library

The Python tools for the Polylith Architecture has support for building libraries to be published at PyPI, even if it isn't the main use case.

Important note: by default, the code in one library will share the same top namespace with other libraries that are built from the same Polylith Monorepo. To solve this, there's a feature available that will organize code according to a custom top namespace and re-write the imports.


You can choose a custom namespace to be used in the build process, by using the --with-top-namespace flag. This is available for Python 3.9 and above.

The build-project command, with a custom top namespace:

poetry build-project --with-top-namespace my_custom_namespace

Hatch, PDM and Rye

A custom top namespace is defined in the project-specific pyproject.toml:

top-namespace = "my_custom_namespace"


By using the Poetry build-project flag or the Hatch Build Hook, the built artifact will look something like this:


And the Python modules will have the custom top namespace as a prefix to imports:

from my_custom_namespace.the_namespace.the_brick import my_function

How is this done?

The command uses AST (Abstract Syntax Tree) parsing to modify source code. The Python built-in ast module is used to parse and un-parse Python code.

Script Entrypoints

If you have a script entrypoint, remember to manually add the custom namespace in your pyproject.toml.

Example (Poetry):

# manually adding the my_custom_namespace at the beginning

my_script = "my_custom_namespace.the_namespace.the_brick.my_module:main"