I'm having a problem retrieving the author metadata information of a package through importlib.metadata.
My package setup is as follows:
I'm using Python's setuptools with a minimal setup.py and some project metadata in pyproject.toml.
setup.py:
from setuptools import setup
setup()
pyproject.toml:
[project]
name = "myproj"
version = "0.1.0"
description = "Some description"
authors = [
{name = "Some One", email = "[email protected]"},
{name = "Another Person", email = "[email protected]"}
]
readme = "README.md"
requires-python = ">=3.11"
license = {text = "BSD-3-Clause"}
[build-system]
requires = ["setuptools"]
build-backend = "setuptools.build_meta"
(I think setup.py is not even necessary for this, but for completeness.)
The pyproject.toml follows keywords described in the Python packaging user guide - Declaring project metadata.
The project itself is very simple: it's structure is
myproj/
__init__.py
and in myproj/__init__.py, I'm using importlib.metadata to get some metadata information about the package.:
from importlib.metadata import metadata
meta = metadata(__package__ or __name__)
print(meta['Name'])
print(meta['Version'])
print(meta['Author-email'])
print(meta['Author'])
I install the package in a virtualenv using pip.
Outside of the project directory, but with the virtualenv activated, I run Python and import the package:
python -c "import myproj"
This prints the various metadata fields:
myproj
0.1.0
Some One <[email protected]>, Another Person <[email protected]>
None
This is where the problem is: the "Author-email" field yields the full [project.authors] entry, as a single string, while the "Author" field yields None.
Is there a way to get the individual authors (without their email) out of the metadata from within the package?
Or is importlib.metadata not up to date with the packaging metadata? (I did a quick search on bugs.python.org, but couldn't find anything related.)
(Of course, I can parse the "Author-email" field manually and separate it into appropriate parts, but it would be nice if that can be avoided.)
Regarding the "author" and "author email" fields in Python packaging metadata:
If I were you I would look into Python's own email library. There is a good chance it has the necessary features to parse the authors metadata properly.
It seems like something like the following could help (modified from this answer):
import email.policy
addresses = 'Some One <[email protected]>, Another Person <[email protected]>'
em = email.message_from_string(
f'To: {addresses}',
policy=email.policy.default,
)
for address in em['to'].addresses:
print(f'{address.display_name} <{address.addr_spec}>')