Look for the well-known project URL keys

According to the official specification:
https://packaging.python.org/en/latest/specifications/well-known-project-urls/
project URLs' labels can be normalized and then searched for the
matches.
Attempt to find "homepage" and "source" first, and only return other
project URL values if these are not found.

pyp2spec/pyp2conf.py

@@ -7,7 +7,7 @@ import tomli_w
 from pyp2spec.license_processor import check_compliance, resolve_license_expression
 from pyp2spec.utils import Pyp2specError, normalize_name, get_extras, get_summary_or_placeholder
 from pyp2spec.utils import prepend_name_with_python, archive_name
-from pyp2spec.utils import is_archful, get_first_url_or_placeholder
+from pyp2spec.utils import is_archful, resolve_url
 from pyp2spec.utils import warn, caution, inform, yay
 from pyp2spec.pypi_loaders import load_from_pypi, load_core_metadata_from_pypi, CoreMetadataNotFoundError
 
@@ -28,7 +28,7 @@ def prepare_package_info(data):
         "pypi_name": normalize_name(data.get("name")),
         "pypi_version": data.get("version"),
         "summary": get_summary_or_placeholder(data.get("summary")),
-        "url": get_first_url_or_placeholder(project_urls),
+        "url": resolve_url(project_urls),
         "extras": get_extras(data.get("requires_dist", [])),
         "license_files_present": bool(data.get("license_files")),
         "license": resolve_license_expression(data),

pyp2spec/utils.py

@@ -3,6 +3,7 @@ This module contains common functions and classes imported in multiple other mod
 """
 
 import re
+import string
 from functools import partial
 
 import click
@@ -152,13 +153,31 @@ def is_archful(archive_urls):
     return False
 
 
-def get_first_url_or_placeholder(urls):
+def _normalize_url_label(label: str) -> str:
+    """
+    Copied verbatim from
+    https://packaging.python.org/en/latest/specifications/well-known-project-urls/#label-normalization
+    """
+    chars_to_remove = string.punctuation + string.whitespace
+    removal_map = str.maketrans("", "", chars_to_remove)
+    return label.translate(removal_map).lower()
+
+
+def resolve_url(urls):
     """
     Project urls are an optional field and come in a form of a dict, e.g.:
     {"homepage": "https://mypackagehomepage.com"}.
     There may be multiple urls defined and keys are arbitrary.
-    As there's no way to programmatically determine "the best" url for the
-    spec file, return the first existing url from the dict.
+    Try to return url for normalized "homepage", then "source" keys.
+    Otherwise return the first value from the dictionary or "..." if the dict is empty.
     """
-
-    return list(urls.values())[0] if urls else "..."
+    normalized_urls = {_normalize_url_label(k): v for k, v in urls.items()}
+    print(normalized_urls)
+    return (
+        normalized_urls.get("homepage")
+        or normalized_urls.get("source")
+        or normalized_urls.get("sourcecode")
+        or normalized_urls.get("github")
+        or normalized_urls.get("repository")
+        or next(iter(normalized_urls.values()), "...")
+    )

tests/expected_specfiles/python-click.spec

@@ -7,7 +7,7 @@ Summary:        Composable command line interface toolkit
 # Check if the automatically generated License and its spelling is correct for Fedora
 # https://docs.fedoraproject.org/en-US/packaging-guidelines/LicensingGuidelines/
 License:        BSD-3-Clause
-URL:            https://palletsprojects.com/donate
+URL:            https://github.com/pallets/click/
 Source:         %{pypi_source click}
 
 BuildArch:      noarch

tests/expected_specfiles/python-numpy.spec

@@ -7,7 +7,7 @@ Summary:        Fundamental package for array computing in Python
 # Check if the automatically generated License and its spelling is correct for Fedora
 # https://docs.fedoraproject.org/en-US/packaging-guidelines/LicensingGuidelines/
 License:        BSD-3-Clause
-URL:            https://github.com/numpy/numpy/issues
+URL:            https://github.com/numpy/numpy
 Source:         %{pypi_source numpy}
 
 BuildRequires:  python3-devel

tests/expected_specfiles/python-sphinx.spec

@@ -6,7 +6,7 @@ Summary:        Python documentation generator
 
 # No license information obtained, it's up to the packager to fill it in
 License:        ...
-URL:            https://www.sphinx-doc.org/en/master/changes.html
+URL:            https://www.sphinx-doc.org/
 Source:         %{pypi_source sphinx}
 
 BuildArch:      noarch

tests/test_configs/customized_python-sphinx.conf

@@ -2,7 +2,7 @@ summary = "Python documentation generator"
 pypi_version = "8.1.3"
 pypi_name = "sphinx"
 python_name = "python-sphinx"
-url = "https://www.sphinx-doc.org/en/master/changes.html"
+url = "https://www.sphinx-doc.org/"
 source = "PyPI"
 extras = [
     "docs",

tests/test_configs/default_python-click.conf

@@ -3,7 +3,7 @@ license = "BSD-3-Clause"
 pypi_name = "click"
 pypi_version = "8.1.7"
 python_name = "python-click"
-url = "https://palletsprojects.com/donate"
+url = "https://github.com/pallets/click/"
 source = "PyPI"
 extras = []
 archful = false

tests/test_configs/default_python-numpy.conf

@@ -4,7 +4,7 @@ pypi_version = "1.25.2"
 license = "BSD-3-Clause"
 pypi_name = "numpy"
 python_name = "python-numpy"
-url = "https://github.com/numpy/numpy/issues"
+url = "https://github.com/numpy/numpy"
 source = "PyPI"
 automode = true
 extras = []

tests/test_utils.py

@@ -3,7 +3,7 @@ import pytest
 from pyp2spec.utils import filter_license_classifiers, prepend_name_with_python
 from pyp2spec.utils import normalize_name, get_extras, is_archful
 from pyp2spec.utils import normalize_as_wheel_name, archive_name
-from pyp2spec.utils import get_first_url_or_placeholder, SdistNotFoundError
+from pyp2spec.utils import resolve_url, SdistNotFoundError
 
 
 def test_license_classifier_read_correctly():
@@ -161,7 +161,7 @@ def test_archive_name_empty_list():
 
 def test_project_urls_valid_single():
     urls = {"homepage": "https://example.com"}
-    assert get_first_url_or_placeholder(urls) == "https://example.com"
+    assert resolve_url(urls) == "https://example.com"
 
 
 def test_project_urls_valid_multiple():
@@ -169,8 +169,8 @@ def test_project_urls_valid_multiple():
         "homepage": "https://example.com",
         "documentation": "https://docs.example.com",
     }
-    assert get_first_url_or_placeholder(urls) == "https://example.com"
+    assert resolve_url(urls) == "https://example.com"
 
 
 def test_project_urls_empty():
-    assert get_first_url_or_placeholder({}) == "..."
+    assert resolve_url({}) == "..."