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.

Packaging

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

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

Deploying

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 plugin has support for building libraries to be published at PyPI, even if it isn't the main use case. But why? 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

By using the --with-top-namespace flag, the built artifact will look something like this:

my_custom_namespace/
    /my_package
       __init__.py
       my_module.py

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

from my_custom_namespace.my_package 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.