Add ability to sync by digests

fixes: #1909
Assisted by: claude-sonnet-4.6
https://redhat.atlassian.net/browse/PULP-1740

Author: gerrod3

CHANGES/+include_exclude_tags.removal

@@ -0,0 +1 @@
+Renamed `include_tags` and `exclude_tags` to `includes` and `excludes` on the remote.

CHANGES/1909.feature

@@ -0,0 +1 @@
+Added support for syncing manifests by digest through new `includes` field on the remote.

pulp_container/app/migrations/0048_containerremote_includes_excludes.py

@@ -0,0 +1,52 @@
+import django.contrib.postgres.fields
+from django.db import migrations, models
+
+
+def migrate_to_includes_excludes(apps, schema_editor):
+    """Copy include_tags -> includes, exclude_tags -> excludes."""
+    ContainerRemote = apps.get_model("container", "ContainerRemote")
+    for remote in ContainerRemote.objects.all():
+        remote.includes = remote.include_tags or None
+        remote.excludes = remote.exclude_tags or None
+        remote.save(update_fields=["includes", "excludes"])
+
+def down_migrate_to_include_exclude_tags(apps, schema_editor):
+    """Copy includes + excludes -> include_tags + exclude_tags."""
+    ContainerRemote = apps.get_model("container", "ContainerRemote")
+    for remote in ContainerRemote.objects.all():
+        remote.include_tags = remote.includes or None
+        remote.exclude_tags = remote.excludes or None
+        remote.save(update_fields=["include_tags", "exclude_tags"])
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("container", "0047_containernamespace_pulp_labels"),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name="containerremote",
+            name="includes",
+            field=django.contrib.postgres.fields.ArrayField(
+                base_field=models.TextField(null=True),
+                null=True,
+                size=None,
+            ),
+        ),
+        migrations.AddField(
+            model_name="containerremote",
+            name="excludes",
+            field=django.contrib.postgres.fields.ArrayField(
+                base_field=models.TextField(null=True),
+                null=True,
+                size=None,
+            ),
+        ),
+        # 2. Copy existing data.
+        migrations.RunPython(
+            migrate_to_includes_excludes,
+            down_migrate_to_include_exclude_tags,
+        ),
+    ]

pulp_container/app/models.py

@@ -483,17 +483,22 @@ class ContainerRemote(Remote, AutoAddObjPermsMixin):
         upstream_name (models.TextField): The name of the image at the remote.
         include_foreign_layers (models.BooleanField): Foreign layers in the remote
             are included. They are not included by default.
-        include_tags (fields.ArrayField): List of tags to include during sync.
-        exclude_tags (fields.ArrayField): List of tags to exclude during sync.
+        includes (fields.ArrayField): List of tags (with optional wildcards) and/or
+            digests (sha256:<hex>) to sync.
+        excludes (fields.ArrayField): List of tag patterns to exclude from sync.
         sigstore (models.TextField): The URL to a sigstore where signatures of container images
             should be synced from.
     """
 
     upstream_name = models.TextField(db_index=True)
     include_foreign_layers = models.BooleanField(default=False)
+    includes = fields.ArrayField(models.TextField(null=True), null=True)
+    excludes = fields.ArrayField(models.TextField(null=True), null=True)
+    sigstore = models.TextField(null=True)
+
+    # Deprecated: kept for ZDT upgrades
     include_tags = fields.ArrayField(models.TextField(null=True), null=True)
     exclude_tags = fields.ArrayField(models.TextField(null=True), null=True)
-    sigstore = models.TextField(null=True)
 
     TYPE = "container"
 

pulp_container/app/serializers.py

@@ -296,27 +296,58 @@ class ContainerRemoteSerializer(RemoteSerializer):
     upstream_name = serializers.CharField(
         required=True, allow_blank=False, help_text=_("Name of the upstream repository")
     )
+    includes = serializers.ListField(
+        child=serializers.CharField(max_length=255),
+        allow_null=True,
+        required=False,
+        help_text=_(
+            "A list of tags (wildcards *, ? are recognized) and/or digests "
+            "(format: 'sha256:<hex>') to include during sync. "
+            "'includes' is evaluated before 'excludes'."
+        ),
+    )
+    excludes = serializers.ListField(
+        child=serializers.CharField(max_length=255),
+        allow_null=True,
+        required=False,
+        help_text=_(
+            "A list of tag patterns to exclude during sync. "
+            "Wildcards *, ? are recognized. "
+            "'excludes' is evaluated after 'includes'."
+        ),
+    )
+
+    # ---------------------------------------------------------------------------
+    # Deprecated write-only fields kept for backwards compatibility.
+    # They are merged into `includes` / `excludes` during validation.
+    # ---------------------------------------------------------------------------
     include_tags = serializers.ListField(
         child=serializers.CharField(max_length=255),
         allow_null=True,
         required=False,
-        help_text=_("""
-            A list of tags to include during sync.
-            Wildcards *, ? are recognized.
-            'include_tags' is evaluated before 'exclude_tags'.
-            """),
+        write_only=True,
+        help_text=_("Deprecated. Use 'includes' instead."),
     )
     exclude_tags = serializers.ListField(
         child=serializers.CharField(max_length=255),
         allow_null=True,
         required=False,
-        help_text=_("""
-            A list of tags to exclude during sync.
-            Wildcards *, ? are recognized.
-            'exclude_tags' is evaluated after 'include_tags'.
-            """),
+        write_only=True,
+        help_text=_("Deprecated. Use 'excludes' instead."),
     )
 
+    def validate(self, data):
+        include_tags = data.pop("include_tags", None)
+        exclude_tags = data.pop("exclude_tags", None)
+
+        if include_tags:
+            data["includes"] = list(data.get("includes") or []) + list(include_tags)
+
+        if exclude_tags:
+            data["excludes"] = list(data.get("excludes") or []) + list(exclude_tags)
+
+        return data
+
     policy = serializers.ChoiceField(
         help_text="""
         immediate - All manifests and blobs are downloaded and saved during a sync.
@@ -337,6 +368,8 @@ class ContainerRemoteSerializer(RemoteSerializer):
     class Meta:
         fields = RemoteSerializer.Meta.fields + (
             "upstream_name",
+            "includes",
+            "excludes",
             "include_tags",
             "exclude_tags",
             "sigstore",

pulp_container/app/tasks/sync_stages.py

@@ -7,6 +7,7 @@
 
 import aiohttp
 from asgiref.sync import sync_to_async
+from collections import defaultdict
 
 from pulpcore.plugin.models import Artifact, ProgressReport, Remote
 from pulpcore.plugin.stages import ContentSaver, DeclarativeArtifact, DeclarativeContent, Stage
@@ -62,7 +63,7 @@ def __init__(self, remote, signed_only):
         self.manifest_list_dcs = []
         self.manifest_dcs = []
         self.signature_dcs = []
-        self._synced_digests = set()
+        self._synced_digests = defaultdict(list)
         self._full_tag_list = []
         self._cosign_tags = []
 
@@ -78,20 +79,26 @@ async def _download_manifest_data(self, manifest_url):
 
         return content_data, raw_text_data, response
 
-    async def _check_for_existing_manifest(self, download_tag):
-        response = await download_tag
+    async def _check_for_existing_manifest(self, head_manifest_task):
+        response = await head_manifest_task
 
         digest = response.headers.get("docker-content-digest")
+        url = response.url
+        original_reference = url.split("/")[-1]
 
         if manifest := await Manifest.objects.filter(
             digest=digest, pulp_domain=get_domain()
         ).afirst():
             raw_text_data = manifest.data
             content_data = json.loads(raw_text_data)
         else:
-            content_data, raw_text_data, response = await self._download_manifest_data(response.url)
+            if not original_reference.startswith("sha256:"):
+                # Fetch the tag with its digest
+                url = url.rsplit(original_reference, 1)[0] + digest
 
-        return content_data, raw_text_data, response
+            content_data, raw_text_data, response = await self._download_manifest_data(url)
+
+        return content_data, raw_text_data, response, original_reference
 
     async def run(self):
         """
@@ -105,31 +112,38 @@ async def run(self):
             repo_name = self.remote.namespaced_upstream_name
             tag_list_url = "/v2/{name}/tags/list".format(name=repo_name)
             self._full_tag_list = await self.get_paginated_tag_list(tag_list_url, repo_name)
-            self._cosign_tags = filter_resources(
-                self._full_tag_list, ["sha256-*"], self.remote.exclude_tags
-            )
-            if self.remote.include_tags or self.remote.exclude_tags:
+            includes = self.remote.includes or []
+            excludes = self.remote.excludes or []
+
+            tag_includes = [i for i in includes if not i.startswith("sha256:")]
+            digest_includes = [i for i in includes if i.startswith("sha256:")]
+            tag_excludes = [e for e in excludes if not e.startswith("sha256:")]
+
+            self._cosign_tags = filter_resources(self._full_tag_list, ["sha256-*"], tag_excludes)
+
+            if includes or excludes:
                 # Split sync into two parts, first all non-cosign tags, then cosign tags
-                exclude_tags_and_cosign = (self.remote.exclude_tags or []) + ["sha256-*"]
-                tag_list = filter_resources(
-                    self._full_tag_list, self.remote.include_tags, exclude_tags_and_cosign
+                exclude_tags_and_cosign = tag_excludes + ["sha256-*"]
+                filtered_tags = filter_resources(
+                    self._full_tag_list, tag_includes, exclude_tags_and_cosign
                 )
+                manifest_list = filtered_tags + digest_includes
             else:
-                tag_list = self._full_tag_list
+                manifest_list = self._full_tag_list
             await pb.aincrement()
 
-        await self._process_tags(tag_list, signature_source)
+        await self._process_manifests(manifest_list, signature_source, "Processing Manifests")
 
-        if self.remote.include_tags or self.remote.exclude_tags:
-            # Process cosign companion tags after all non-cosign tags are synced
+        if includes or excludes:
+            # Process cosign companion tags after all primary content is synced
             companion_tags = self._find_cosign_companion_tags()
             if companion_tags:
                 log.info(
                     "Syncing %d cosign companion tag(s) for filtered images",
                     len(companion_tags),
                 )
-                await self._process_tags(
-                    companion_tags, signature_source, msg="Processing Cosign Companion Tags"
+                await self._process_manifests(
+                    companion_tags, signature_source, "Processing Cosign Companion Tags"
                 )
 
     def _find_cosign_companion_tags(self):
@@ -143,54 +157,53 @@ def _find_cosign_companion_tags(self):
                 companion_tags.append(tag)
         return companion_tags
 
-    async def _process_tags(self, tag_list, signature_source, msg="Processing Tags"):
-        """Download and process a batch of tags, creating declarative content objects."""
+    async def _process_manifests(self, manifests, signature_source, msg):
+        """Download and process a batch of manifests, creating declarative content objects."""
         BATCH_SIZE = 500
         to_download = []
 
-        for tag_name in tag_list:
-            relative_url = "/v2/{name}/manifests/{tag}".format(
-                name=self.remote.namespaced_upstream_name, tag=tag_name
-            )
-            tag_url = urljoin(self.remote.url, relative_url)
-            downloader = self.remote.get_downloader(url=tag_url)
+        for reference in manifests:
+            relative_url = f"/v2/{self.remote.namespaced_upstream_name}/manifests/{reference}"
+            manifest_url = urljoin(self.remote.url, relative_url)
+            downloader = self.remote.get_downloader(url=manifest_url)
             to_download.append(
                 downloader.run(extra_data={"headers": V2_ACCEPT_HEADERS, "http_method": "head"})
             )
 
         async with ProgressReport(
             message=msg,
-            code="sync.processing.tag",
-            total=len(tag_list),
-        ) as pb_parsed_tags:
+            code="sync.processing.manifest",
+            total=len(manifests),
+        ) as pb_parsed_manifests:
             to_download_artifact = [
-                self._check_for_existing_manifest(download_tag)
-                for download_tag in asyncio.as_completed(to_download)
+                self._check_for_existing_manifest(download_manifest)
+                for download_manifest in asyncio.as_completed(to_download)
             ]
 
             for artifact in asyncio.as_completed(to_download_artifact):
-                content_data, raw_text_data, response = await artifact
+                content_data, raw_text_data, response, manifest_ref = await artifact
 
                 digest = calculate_digest(raw_text_data)
-                tag_name = response.url.split("/")[-1]
+                is_tag = not manifest_ref.startswith("sha256:")
                 media_type = determine_media_type(content_data, response)
 
                 if self.signed_only and not signature_source:
                     if not (
-                        self._is_cosign_companion_tag(tag_name, media_type, content_data)
+                        self._is_cosign_companion_tag(manifest_ref, media_type, content_data)
                         or await self._has_cosign_signature(digest)
                     ):
                         log.info(
                             "The unsigned image {digest} can't be synced "
                             "due to a requirement to sync signed content "
                             "only.".format(digest=digest)
                         )
-                        await pb_parsed_tags.aincrement()
+                        await pb_parsed_manifests.aincrement()
                         continue
 
                 validate_manifest(content_data, media_type, digest)
 
-                tag_dc = DeclarativeContent(Tag(name=tag_name))
+                if is_tag:
+                    tag_dc = DeclarativeContent(Tag(name=manifest_ref))
 
                 if media_type in (MEDIA_TYPE.MANIFEST_LIST, MEDIA_TYPE.INDEX_OCI):
                     list_dc = self.create_manifest_list(
@@ -214,7 +227,7 @@ async def _process_tags(self, tag_list, signature_source, msg="Processing Tags")
                                     "The whole manifest list is skipped.".format(
                                         img_digest=man_dc.content.digest,
                                         ml_digest=list_dc.content.digest,
-                                        tag=tag_name,
+                                        tag=manifest_ref,
                                     )
                                 )
                                 break
@@ -224,20 +237,23 @@ async def _process_tags(self, tag_list, signature_source, msg="Processing Tags")
                     else:
                         # Manifest indices can be signed too. It is not mandatory.
                         # If signature is available mirror it.
-                        self._synced_digests.add(digest)
+                        self._synced_digests[digest].append(manifest_ref)
                         if signature_source is not None:
                             list_sig_dcs = await self.create_signatures(list_dc, signature_source)
                             if list_sig_dcs:
                                 self.signature_dcs.extend(list_sig_dcs)
-                        tag_dc.extra_data["tagged_manifest_dc"] = list_dc
                         for listed_manifest in list_dc.extra_data["listed_manifests"]:
-                            self._synced_digests.add(listed_manifest["manifest_dc"].content.digest)
+                            self._synced_digests[
+                                listed_manifest["manifest_dc"].content.digest
+                            ].append(manifest_ref)
                             await self.handle_blobs(
                                 listed_manifest["manifest_dc"], listed_manifest["content_data"]
                             )
                             self.manifest_dcs.append(listed_manifest["manifest_dc"])
                         self.manifest_list_dcs.append(list_dc)
-                        self.tag_dcs.append(tag_dc)
+                        if is_tag:
+                            tag_dc.extra_data["tagged_manifest_dc"] = list_dc
+                            self.tag_dcs.append(tag_dc)
 
                 else:
                     # Simple tagged manifest
@@ -249,14 +265,15 @@ async def _process_tags(self, tag_list, signature_source, msg="Processing Tags")
                         if self.signed_only and not man_sig_dcs:
                             continue
                         self.signature_dcs.extend(man_sig_dcs)
-                    self._synced_digests.add(digest)
-                    tag_dc.extra_data["tagged_manifest_dc"] = man_dc
+                    self._synced_digests[digest].append(manifest_ref)
                     await self.handle_blobs(man_dc, content_data)
-                    self.tag_dcs.append(tag_dc)
+                    if is_tag:
+                        tag_dc.extra_data["tagged_manifest_dc"] = man_dc
+                        self.tag_dcs.append(tag_dc)
                     self.manifest_dcs.append(man_dc)
 
                 # Count the skipped tasks as parsed too.
-                await pb_parsed_tags.aincrement()
+                await pb_parsed_manifests.aincrement()
 
                 # Flush the queues to prevent overly excessive memory usage.
                 # This will cap the number of in flight high level objects to about BATCH_SIZE.