This section of the tutorial deals with setup of the project and structure for a Python Polylith project.
We will use the Polylith architecture and Python tooling for it to test how how well the Polylith methods and tools work in a microservices use case.
Create a new empty directory and repo for this project
git init
Install Poetry if not already installed - I used homebrew on Mac to install Poetry
Add the basic poetry setup and follow the instructions
poetry init
Add the poetry polylith and multiproject plugins that will allow us to set up and manage the Polylith setup using poetry
poetry self add poetry-multiproject-plugin
poetry self add poetry-polylith-plugin
Configure Poetry to create virtual environments in a local .venv folder
poetry config virtualenvs.in-project true
For my project I know I am going to need fastapi for the api, and pika for the message queues for this demo. I will also use faker to create dummy data.
Install the dependencies using
poetry add fastapi uvicorn pika faker
Also install the development dependencies you want for your setup. For my setup I like the following with VSCode:
poetry add --dev pylint isort pytest black ipykernel
Polylith has a specific directory structure that separates out the components, bases, projects and development. This is set up in what they call a workspace. The details can be read about in the Polylith documentation.
David Vujic has reinterpreted these rules in the Python setting which is called the loose theme. This uses a directory structure that is more Pythonic and natural to Python devs. This tutorial will follow his suggestions and use this loose theme.
We use poetry with the poly plugin to set up our workspace and directory structure. I will give this project a name mousetrap
poetry poly create workspace --name mousetrap --theme loose
This creates a workspace.toml file and also the directories for components, bases, projects and development.
We are now ready to start writing some code.