First published release of Flower!

We’ve published our third, core Python package :tada: :grin: This package turns the machine-readable and structured metadata from a data package into a format that is easier for humans to view and use.

On March 25th, 2026, we published our third Python package, Flower, to PyPI. This package converts a data package’s structured, machine-readable metadata to something more suited to humans, Markdown files, that can be displayed on e.g., a website. This is our first package that non-data engineers might interact with and use (compared to our other two packages, Sprout or check-datapackge). The output of Flower is (ideally) one of the first ways that researchers interact with a data package from a research study.

What is Flower?

Flower’s dedicated website has more details about it, but in brief, Flower is a command-line tool that takes a metadata file, like datapackage.json, and converts it into human-readable Markdown files. These Markdown files can then be converted into HTML or PDF files for easier sharing with researchers or others who may use the data in the data package. Our aim is that Flower will help with the issue that researchers frequently face when working with data: the lack of documentation about the data they’re working with.

Why use it?

Aligning with our general principles of keeping tools simple and focused, we kept Flower sharply focused on output a human-readable document from machine-readable metadata. It’s also highly customisable, so if our current styles don’t fit your needs, you aren’t restricted to using those. We will also continue to add more styles and improve on the existing ones.

If you want to see what some of those styles look like, check out our “Examples” page.

Commands in Flower

There are (so far) two commands in Flower: build and view. You can run them fairly easily with uv.

To see the help docs for them, run:

Terminal

uvx seedcase-flower

To build the metadata into something human-friendly, run:

Terminal

uvx seedcase-flower build --help

To quickly view the metadata (e.g., if the owner of the data package hasn’t yet made a website but have structured the metadata into a metadata file like datapackage.json), run:

Terminal

uvx seedcase-flower view --help

We do have plans to make an interactive terminal-based viewer for those who prefer working in the terminal.