Merge pull request #885 from jobselko/metadata_support

Fully support the latest core metadata (up to 2.4)

Author: jobselko

CHANGES/689.feature

@@ -0,0 +1 @@
+Added full support for the latest core metadata (up to 2.4).

pulp_python/app/migrations/0014_pythonpackagecontent_dynamic_and_more.py

@@ -0,0 +1,34 @@
+# Generated by Django 4.2.19 on 2025-07-09 08:05
+
+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),
+        ),
+        migrations.AddField(
+            model_name='pythonpackagecontent',
+            name='license_expression',
+            field=models.TextField(default=''),
+            preserve_default=False,
+        ),
+        migrations.AddField(
+            model_name='pythonpackagecontent',
+            name='license_file',
+            field=models.JSONField(default=list),
+        ),
+        migrations.AddField(
+            model_name='pythonpackagecontent',
+            name='provides_extras',
+            field=models.JSONField(default=list),
+        ),
+    ]

pulp_python/app/models.py

@@ -140,49 +140,66 @@ class PythonPackageContent(Content):
     """
     A Content Type representing Python's Distribution Package.
 
-    As defined in pep-0426 and pep-0345.
+    Core Metadata:
+        https://packaging.python.org/en/latest/specifications/core-metadata/
 
-    https://www.python.org/dev/peps/pep-0491/
-    https://www.python.org/dev/peps/pep-0345/
-    """
-
-    PROTECTED_FROM_RECLAIM = False
+    Release metadata (JSON API):
+        https://docs.pypi.org/api/json/
 
-    TYPE = "python"
-    repo_key_fields = ("filename",)
-    # Required metadata
-    filename = models.TextField(db_index=True)
-    packagetype = models.TextField(choices=PACKAGE_TYPES)
-    name = models.TextField()
-    name.register_lookup(NormalizeName)
-    version = models.TextField()
-    sha256 = models.CharField(db_index=True, max_length=64)
-    # Optional metadata
-    python_version = models.TextField()
-    metadata_version = models.TextField()
-    summary = models.TextField()
-    description = models.TextField()
-    keywords = models.TextField()
-    home_page = models.TextField()
-    download_url = models.TextField()
+    File Formats:
+        https://packaging.python.org/en/latest/specifications/source-distribution-format/
+        https://packaging.python.org/en/latest/specifications/binary-distribution-format/
+    """
+    # Core metadata
+    # Version 1.0
     author = models.TextField()
     author_email = models.TextField()
+    description = models.TextField()
+    home_page = models.TextField()  # Deprecated in favour of Project-URL
+    keywords = models.TextField()
+    license = models.TextField()  # Deprecated in favour of License-Expression
+    metadata_version = models.TextField()
+    name = models.TextField()
+    platform = models.TextField()
+    summary = models.TextField()
+    version = models.TextField()
+    # Version 1.1
+    classifiers = models.JSONField(default=list)
+    download_url = models.TextField()  # Deprecated in favour of Project-URL
+    supported_platform = models.TextField()
+    # Version 1.2
     maintainer = models.TextField()
     maintainer_email = models.TextField()
-    license = models.TextField()
-    requires_python = models.TextField()
+    obsoletes_dist = models.JSONField(default=list)
     project_url = models.TextField()
-    platform = models.TextField()
-    supported_platform = models.TextField()
-    requires_dist = models.JSONField(default=list)
+    project_urls = models.JSONField(default=dict)
     provides_dist = models.JSONField(default=list)
-    obsoletes_dist = models.JSONField(default=list)
     requires_external = models.JSONField(default=list)
-    classifiers = models.JSONField(default=list)
-    project_urls = models.JSONField(default=dict)
+    requires_dist = models.JSONField(default=list)
+    requires_python = models.TextField()
+    # Version 2.1
     description_content_type = models.TextField()
-    # Pulp Domains
-    _pulp_domain = models.ForeignKey("core.Domain", default=get_domain_pk, on_delete=models.PROTECT)
+    provides_extras = models.JSONField(default=list)
+    # Version 2.2
+    dynamic = models.JSONField(default=list)
+    # Version 2.4
+    license_expression = models.TextField()
+    license_file = models.JSONField(default=list)
+
+    # Release metadata
+    filename = models.TextField(db_index=True)
+    packagetype = models.TextField(choices=PACKAGE_TYPES)
+    python_version = models.TextField()
+    sha256 = models.CharField(db_index=True, max_length=64)
+
+    # From pulpcore
+    PROTECTED_FROM_RECLAIM = False
+    TYPE = "python"
+    _pulp_domain = models.ForeignKey(
+        "core.Domain", default=get_domain_pk, on_delete=models.PROTECT
+    )
+    name.register_lookup(NormalizeName)
+    repo_key_fields = ("filename",)
 
     @staticmethod
     def init_from_artifact_and_relative_path(artifact, relative_path):

pulp_python/app/serializers.py

@@ -72,69 +72,69 @@ class PythonPackageContentSerializer(core_serializers.SingleArtifactContentUploa
     """
     A Serializer for PythonPackageContent.
     """
-
-    filename = serializers.CharField(
-        help_text=_('The name of the distribution package, usually of the format:'
-                    ' {distribution}-{version}(-{build tag})?-{python tag}-{abi tag}'
-                    '-{platform tag}.{packagetype}'),
-        read_only=True,
-    )
-    packagetype = serializers.CharField(
-        help_text=_('The type of the distribution package '
-                    '(e.g. sdist, bdist_wheel, bdist_egg, etc)'),
-        read_only=True,
-    )
-    name = serializers.CharField(
-        help_text=_('The name of the python project.'),
-        read_only=True,
-    )
-    version = serializers.CharField(
-        help_text=_('The packages version number.'),
-        read_only=True,
-    )
-    sha256 = serializers.CharField(
-        default='',
-        help_text=_('The SHA256 digest of this package.'),
-    )
-    metadata_version = serializers.CharField(
-        help_text=_('Version of the file format'),
-        read_only=True,
+    # Core metadata
+    # Version 1.0
+    author = serializers.CharField(
+        required=False, allow_blank=True,
+        help_text=_('Text containing the author\'s name. Contact information can also be added,'
+                    ' separated with newlines.')
     )
-    summary = serializers.CharField(
+    author_email = serializers.CharField(
         required=False, allow_blank=True,
-        help_text=_('A one-line summary of what the package does.')
+        help_text=_('The author\'s e-mail address. ')
     )
     description = serializers.CharField(
         required=False, allow_blank=True,
         help_text=_('A longer description of the package that can run to several paragraphs.')
     )
-    description_content_type = serializers.CharField(
+    home_page = 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.')
+        help_text=_('The URL for the package\'s home page.')
     )
     keywords = serializers.CharField(
         required=False, allow_blank=True,
         help_text=_('Additional keywords to be used to assist searching for the '
                     'package in a larger catalog.')
     )
-    home_page = serializers.CharField(
+    license = serializers.CharField(
         required=False, allow_blank=True,
-        help_text=_('The URL for the package\'s home page.')
+        help_text=_('Text indicating the license covering the distribution')
     )
-    download_url = serializers.CharField(
+    metadata_version = serializers.CharField(
+        help_text=_('Version of the file format'),
+        read_only=True,
+    )
+    name = serializers.CharField(
+        help_text=_('The name of the python project.'),
+        read_only=True,
+    )
+    platform = serializers.CharField(
         required=False, allow_blank=True,
-        help_text=_('Legacy field denoting the URL from which this package can be downloaded.')
+        help_text=_('A comma-separated list of platform specifications, '
+                    'summarizing the operating systems supported by the package.')
     )
-    author = serializers.CharField(
+    summary = serializers.CharField(
         required=False, allow_blank=True,
-        help_text=_('Text containing the author\'s name. Contact information can also be added,'
-                    ' separated with newlines.')
+        help_text=_('A one-line summary of what the package does.')
     )
-    author_email = serializers.CharField(
+    version = serializers.CharField(
+        help_text=_('The packages version number.'),
+        read_only=True,
+    )
+    # Version 1.1
+    classifiers = serializers.JSONField(
+        required=False, default=list,
+        help_text=_('A JSON list containing classification values for a Python package.')
+    )
+    download_url = serializers.CharField(
         required=False, allow_blank=True,
-        help_text=_('The author\'s e-mail address. ')
+        help_text=_('Legacy field denoting the URL from which this package can be downloaded.')
     )
+    supported_platform = serializers.CharField(
+        required=False, allow_blank=True,
+        help_text=_('Field to specify the OS and CPU for which the binary package was compiled. ')
+    )
+    # Version 1.2
     maintainer = serializers.CharField(
         required=False, allow_blank=True,
         help_text=_('The maintainer\'s name at a minimum; '
@@ -144,14 +144,11 @@ class PythonPackageContentSerializer(core_serializers.SingleArtifactContentUploa
         required=False, allow_blank=True,
         help_text=_('The maintainer\'s e-mail address.')
     )
-    license = serializers.CharField(
-        required=False, allow_blank=True,
-        help_text=_('Text indicating the license covering the distribution')
-    )
-    requires_python = serializers.CharField(
-        required=False, allow_blank=True,
-        help_text=_('The Python version(s) that the distribution is guaranteed to be '
-                    'compatible with.')
+    obsoletes_dist = serializers.JSONField(
+        required=False, default=list,
+        help_text=_('A JSON list containing names of a distutils project\'s distribution which '
+                    'this distribution renders obsolete, meaning that the two projects should not '
+                    'be installed at the same time.')
     )
     project_url = serializers.CharField(
         required=False, allow_blank=True,
@@ -161,39 +158,73 @@ class PythonPackageContentSerializer(core_serializers.SingleArtifactContentUploa
         required=False, default=dict,
         help_text=_('A dictionary of labels and URLs for the project.')
     )
-    platform = serializers.CharField(
-        required=False, allow_blank=True,
-        help_text=_('A comma-separated list of platform specifications, '
-                    'summarizing the operating systems supported by the package.')
+    provides_dist = serializers.JSONField(
+        required=False, default=list,
+        help_text=_('A JSON list containing names of a Distutils project which is contained'
+                    ' within this distribution.')
     )
-    supported_platform = serializers.CharField(
-        required=False, allow_blank=True,
-        help_text=_('Field to specify the OS and CPU for which the binary package was compiled. ')
+    requires_external = serializers.JSONField(
+        required=False, default=list,
+        help_text=_('A JSON list containing some dependency in the system that the distribution '
+                    'is to be used.')
     )
     requires_dist = serializers.JSONField(
         required=False, default=list,
         help_text=_('A JSON list containing names of some other distutils project '
                     'required by this distribution.')
     )
-    provides_dist = serializers.JSONField(
-        required=False, default=list,
-        help_text=_('A JSON list containing names of a Distutils project which is contained'
-                    ' within this distribution.')
+    requires_python = serializers.CharField(
+        required=False, allow_blank=True,
+        help_text=_('The Python version(s) that the distribution is guaranteed to be '
+                    'compatible with.')
     )
-    obsoletes_dist = serializers.JSONField(
+    # Version 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_extras = serializers.JSONField(
         required=False, default=list,
-        help_text=_('A JSON list containing names of a distutils project\'s distribution which '
-                    'this distribution renders obsolete, meaning that the two projects should not '
-                    'be installed at the same time.')
+        help_text=_('A JSON list containing names of optional features provided by the package.')
     )
-    requires_external = serializers.JSONField(
+    # Version 2.2
+    dynamic = serializers.JSONField(
         required=False, default=list,
-        help_text=_('A JSON list containing some dependency in the system that the distribution '
-                    'is to be used.')
+        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.')
     )
-    classifiers = serializers.JSONField(
+    # Version 2.4
+    license_expression = serializers.CharField(
+        required=False, allow_blank=True,
+        help_text=_('Text string that is a valid SPDX license expression.')
+    )
+    license_file = serializers.JSONField(
         required=False, default=list,
-        help_text=_('A JSON list containing classification values for a Python package.')
+        help_text=_('A JSON list containing names of the paths to license-related files.')
+    )
+    # Release metadata
+    filename = serializers.CharField(
+        help_text=_('The name of the distribution package, usually of the format:'
+                    ' {distribution}-{version}(-{build tag})?-{python tag}-{abi tag}'
+                    '-{platform tag}.{packagetype}'),
+        read_only=True,
+    )
+    packagetype = serializers.CharField(
+        help_text=_('The type of the distribution package '
+                    '(e.g. sdist, bdist_wheel, bdist_egg, etc)'),
+        read_only=True,
+    )
+    python_version = serializers.CharField(
+        help_text=_(
+            'The tag that indicates which Python implementation or version the package requires.'
+        ),
+        read_only=True,
+    )
+    sha256 = serializers.CharField(
+        default='',
+        help_text=_('The SHA256 digest of this package.'),
     )
 
     def deferred_validate(self, data):
@@ -242,11 +273,13 @@ 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'
+            'author', 'author_email', 'description', 'home_page', 'keywords', 'license',
+            'metadata_version', 'name', 'platform', 'summary', 'version', 'classifiers',
+            'download_url', 'supported_platform', 'maintainer', 'maintainer_email',
+            'obsoletes_dist', 'project_url', 'project_urls', 'provides_dist', 'requires_external',
+            'requires_dist', 'requires_python', 'description_content_type',
+            'provides_extras', 'dynamic', 'license_expression', 'license_file',
+            'filename', 'packagetype', 'python_version', 'sha256'
         )
         model = python_models.PythonPackageContent