Dense Analysis has released dense-python, a template repository designed to help developers spin up new Python projects with minimal setup. The repository collects a set of best practices used across Dense Analysis projects and packages them into a starting point you can clone and modify.
Available on GitHub, the template comes preconfigured with modern tooling for formatting, linting, static type checks, and continuous integration. By cloning the repository and customizing it for your work, you can skip much of the boilerplate configuration typically required when starting from scratch.
Distroless Python containers
For containerized deployments, dense-python includes a reproducible build
pipeline that produces distroless images. Distroless containers ship only the
Python runtime and your project code, resulting in smaller and more secure
images than traditional distributions. The provided Dockerfile demonstrates
how to compile wheels with the uv build tool and copy them into the final
distroless image. Read more about the advantages in this article on distroless
Python images.
Testing with Network Isolation By Default
The template configures pytest so that tests run with socket access disabled
by default. This makes it easy to catch accidental network calls in unit tests.
Strict marker checking is also enabled, ensuring that every custom marker is
declared in your configuration. You can decorate tests with
@pytest.mark.allow_network to enable network access for them explictly.
Together, these settings help keep tests fast and deterministic.
Pyright Type Checking
Static type checking is handled by
Pyright, the engine behind
Pylance
in VSCode and also used by Cursor. It offers quick
feedback on type annotations without slowing down development. A ready-to-use
pyproject.toml configuration is included so that type errors are caught early.
As ever Dense Analysis recommends using the
DAFT for programming over proprietary
editors, but they may also be easily configured.
Package Builds with uv
Instead of naked pip, dense-python relies on the lightning-fast
uv tool to create virtual environments and
build distribution packages. This speeds up the packaging process and ensures
consistent builds across platforms.
Open for Contributions!
dense-python is FOSS and welcomes contributions. If you find a useful
improvement or fix, open a pull request or start a discussion on the project
page. We hope this template saves you time on setup so you can focus on writing
code.