Update pkginfo library to support newer python packages

closes #689

CHANGES/689.feature

@@ -0,0 +1 @@
+Support Python package metadata version 2.3

pulp_python/app/migrations/0014_pythonpackagecontent_dynamic_and_more.py

@@ -0,0 +1,23 @@
+# Generated by Django 4.2.10 on 2024-07-08 00:47
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('python', '0013_add_rbac_permissions'),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name='pythonpackagecontent',
+            name='dynamic',
+            field=models.JSONField(default=list, null=True),
+        ),
+        migrations.AddField(
+            model_name='pythonpackagecontent',
+            name='provides_extra',
+            field=models.JSONField(default=list, null=True),
+        ),
+    ]

pulp_python/app/models.py

@@ -145,6 +145,10 @@ class PythonPackageContent(Content):
 
     PROTECTED_FROM_RECLAIM = False
 
+    # TODO: it appears we've set the default (usually empty-string) for each of these fields
+    # manually in the migrations rather than setting them declaratively. That's not ideal.
+    # At some point we should add proper default values and probably make some fields nullable.
+
     TYPE = "python"
     repo_key_fields = ("filename",)
     # Required metadata
@@ -177,7 +181,11 @@ class PythonPackageContent(Content):
     requires_external = models.JSONField(default=list)
     classifiers = models.JSONField(default=list)
     project_urls = models.JSONField(default=dict)
+    # Metadata 2.1
     description_content_type = models.TextField()
+    provides_extra = models.JSONField(default=list, null=True)
+    # Metadata 2.2
+    dynamic = models.JSONField(default=list, null=True)
     # Pulp Domains
     _pulp_domain = models.ForeignKey("core.Domain", default=get_domain_pk, on_delete=models.PROTECT)
 

pulp_python/app/serializers.py

@@ -108,11 +108,6 @@ class PythonPackageContentSerializer(core_serializers.SingleArtifactContentUploa
         required=False, allow_blank=True,
         help_text=_('A longer description of the package that can run to several paragraphs.')
     )
-    description_content_type = serializers.CharField(
-        required=False, allow_blank=True,
-        help_text=_('A string stating the markup syntax (if any) used in the distribution’s'
-                    ' description, so that tools can intelligently render the description.')
-    )
     keywords = serializers.CharField(
         required=False, allow_blank=True,
         help_text=_('Additional keywords to be used to assist searching for the '
@@ -195,6 +190,23 @@ class PythonPackageContentSerializer(core_serializers.SingleArtifactContentUploa
         required=False, default=list,
         help_text=_('A JSON list containing classification values for a Python package.')
     )
+    # Metadata 2.1
+    description_content_type = serializers.CharField(
+        required=False, allow_blank=True,
+        help_text=_('A string stating the markup syntax (if any) used in the distribution’s'
+                    ' description, so that tools can intelligently render the description.')
+    )
+    provides_extra = serializers.JSONField(
+        required=False, default=list,
+        help_text=_('A JSON list containing names of optional features provided by the package.')
+    )
+    # Metadata 2.2
+    dynamic = serializers.JSONField(
+        required=False, default=list,
+        help_text=_('A JSON list containing names of other core metadata fields which are '
+                    'permitted to vary between sdist and bdist packages. Fields NOT marked '
+                    'dynamic MUST be the same between bdist and sdist.')
+    )
 
     def deferred_validate(self, data):
         """
@@ -251,10 +263,11 @@ def retrieve(self, validated_data):
     class Meta:
         fields = core_serializers.SingleArtifactContentUploadSerializer.Meta.fields + (
             'filename', 'packagetype', 'name', 'version', 'sha256', 'metadata_version', 'summary',
-            'description', 'description_content_type', 'keywords', 'home_page', 'download_url',
-            'author', 'author_email', 'maintainer', 'maintainer_email', 'license',
-            'requires_python', 'project_url', 'project_urls', 'platform', 'supported_platform',
-            'requires_dist', 'provides_dist', 'obsoletes_dist', 'requires_external', 'classifiers'
+            'description', 'keywords', 'home_page', 'download_url', 'author', 'author_email',
+            'maintainer', 'maintainer_email', 'license', 'requires_python', 'project_url',
+            'project_urls', 'platform', 'supported_platform', 'requires_dist', 'provides_dist',
+            'obsoletes_dist', 'requires_external', 'classifiers', 'description_content_type',
+            'provides_extra', 'dynamic',
         )
         model = python_models.PythonPackageContent
 

pulp_python/app/utils.py

@@ -91,8 +91,9 @@ def parse_project_metadata(project):
     package['obsoletes_dist'] = json.dumps(project.get('obsoletes_dist', []))
     package['requires_external'] = json.dumps(project.get('requires_external', []))
     package['classifiers'] = json.dumps(project.get('classifiers', []))
-    package['project_urls'] = json.dumps(project.get('project_urls', {}))
     package['description_content_type'] = project.get('description_content_type') or ""
+    package['provides_extra'] = project.get('provides_extra', None)
+    package['dynamic'] = project.get('dynamic') or None
 
     return package
 
@@ -219,7 +220,6 @@ def python_content_to_info(content):
         "summary": content.summary or "",
         "keywords": content.keywords or "",
         "description": content.description or "",
-        "description_content_type": content.description_content_type or "",
         "bugtrack_url": None,  # These two are basically never used
         "docs_url": None,
         "downloads": {"last_day": -1, "last_month": -1, "last_week": -1},
@@ -240,6 +240,9 @@ def python_content_to_info(content):
         "classifiers": json_to_dict(content.classifiers) or None,
         "yanked": False,  # These are no longer used on PyPI, but are still present
         "yanked_reason": None,
+        "description_content_type": content.description_content_type or "",
+        "provides_extra": content.provides_extra,
+        "dynamic": content.dynamic,
     }