Packaging¶

Dependency management and packaging for this project is managed by Poetry.

Poetry¶

Poetry is a tool for dependency management and packaging in Python. It allows you to declare the libraries your project depends on and it will manage (install/update) them for you. Poetry offers a lockfile to ensure repeatable installs, and can build your project for distribution.¹

Why Poetry?¶

Personally, I find poetry dependable and easy to use. Of course, there are alternatives to poetry. Examples of these are:

• PDM
• Rye

Using a dependency management tool allows us to easily set up repeatable installations (through a lockfile, such as poetry.lock), as well as package distribution.

The entire package configuration is managed through a single pyproject.toml file located in the repository's root. This is a big improvement compared to the clunky procedure using a classic approach with a setuptools backend.

Moreover, poetry allows easy distinguishing between core, development and test dependencies. This enables the user to choose what to install in an easy manner using the pip CLI.

Drawbacks¶

A principal drawback I find, is that the .venv (Virtual Environment) folder created by Poetry can become quite large.

For projects that require complex dependencies (e.g., PyTorch projects, which ship with GPU/CUDA drivers, etc.), I often use micromamba to create a specific conda environment for the project. This way, I can re-use the same environment for multiple projects.

micromamba is a tiny version of the mamba package manager. It is a statically linked C++ executable with a separate command line interface. It does not need a base environment and does not come with a default version of Python.²

These environments can be persisted in a YAML file to create a functionality similar to a lockfile.