Feature Request: Ability to get any of package's metadata fields from pyproject.toml (e.g. author)

[MarximusMaximus]

I find myself wanting to fill out the rest of the dunder metadata variables for my package, but do not want to maintain the data in two places. I think(?) single-source is the correct answer, but it currently only handles version numbers.

[rabbit72]

hi @MarximusMaximus, thanks for your interest.
What kind of metadata from pyproject.toml are you interested in?

Probably you have some ideas of what code can look like. Code snippets can be very helpful.

[MarximusMaximus]

I'd want the below fields in my package to be pulled from the field from pyproject.toml in parentheses and the importlib.metadata field in square brackets. This would be possible /just/ with using importlib.metadata but that fails if the project is being run directly as source from a git clone.

__author__ (authors) [Author + Author-email]
__authors__ (authors) [Author + Author-email]
__maintainer__ (maintainers) [Maintainer + Maintainer-email]
__maintainers__ (maintainers) [Maintainer + Maintainer-email]
__contact__ (maintainers) [Maintainer + Maintainer-email]
__license__ (license) [License]
__version__ (version) [Version]

maybe some others, but they seem a little less standardized in their dunder var naming like homepage and other urls

(also, for some reason, right now my test files via poetry are only grabbing the first value for list fields like author and urls...probably pebkac though)

[MarximusMaximus]

Adding some references:
dunder metadata vars: https://stackoverflow.com/a/1523456
poetry pyproject.toml vars: https://python-poetry.org/docs/pyproject/
importlib.metadata vars: https://peps.python.org/pep-0345/

note to self: how to properly access multi-specify fields (arrays):

importlib.metadata.metadata("pytest_prefer_nested_dup_tests").get_all("Project-URL")