Tech debt is overdue to be paid off -- what issues are most urgent?

[monocongo]

I have some PTO coming up, and during that time I will attempt to get this repository back into a semblance of order. Some of the tech debts I plan to address:

• get the version in PyPI and the master branch in sync (including removing the version designation from init.py or anywhere else other than in pyproject.toml, see Need to update __version__ in __init__.py #498, Delete setup.py, don't keep a fake version #511), this will hopefully resolve some of the L-moments issues that we thought were fixed in previous releases (maybe we need to also make proper releases and tags?)
• get rid of the various setuptools-related packaging artifacts and shift to using pyproject.toml with poetry for builds, dependency management, etc. (does this include dropping requirements.txt etc.? I'm still learning the best practices around all that.)
• provide sufficient/additional documentation around installation and usage examples to allow for a review of the project for PyOpenSci (see Add API docs and install instructions to README #521)
• re-do the CI/CD process by leveraging GitHub Actions -- the previous setup I cobbled together for Travis, CircleCI etc. is a real mess now, and we want to have a simple setup where any push to master or dev branches causes a run of the unit tests, with this as a requirement for any PR merge

What other outstanding issues does the community want to see addressed, assuming I have enough time to knock out all of the above? Bear in mind that I'm a software engineer and not a scientist, so working out any of the actual scientific algorithms etc. will be out of scope, I'm just trying to provide a solid package for scientific programmers that follows "best practices".

Thanks in advance, as always, to the users and contributors of this package. I hope it continues to be useful.