Read blobs from new manifest blob table where relevant

data/model/oci/__init__.py

@@ -1,6 +1,7 @@
 # There MUST NOT be any circular dependencies between these subsections. If there are fix it by
 # moving the minimal number of things to shared
 from data.model.oci import (
+  blob,
   label,
   manifest,
   shared,

data/model/oci/blob.py

@@ -0,0 +1,26 @@
+from data.database import ImageStorage, ManifestBlob
+from data.model import BlobDoesNotExist
+from data.model.storage import get_storage_by_subquery, InvalidImageException
+from data.model.blob import get_repository_blob_by_digest as legacy_get
+
+def get_repository_blob_by_digest(repository, blob_digest):
+  """ Find the content-addressable blob linked to the specified repository and
+      returns it or None if none.
+  """
+  try:
+    storage_id_query = (ImageStorage
+                        .select(ImageStorage.id)
+                        .join(ManifestBlob)
+                        .where(ManifestBlob.repository == repository,
+                               ImageStorage.content_checksum == blob_digest,
+                               ImageStorage.uploading == False)
+                        .limit(1))
+
+    return get_storage_by_subquery(storage_id_query)
+  except InvalidImageException:
+    # TODO(jschorr): Remove once we are no longer using the legacy tables.
+    # Try the legacy call.
+    try:
+      return legacy_get(repository, blob_digest)
+    except BlobDoesNotExist:
+      return None

data/model/storage.py

@@ -270,7 +270,7 @@ def get_layer_path_for_storage(storage_uuid, cas_path, content_checksum):
   return store.blob_path(content_checksum)
 
 
-def lookup_repo_storages_by_content_checksum(repo, checksums):
+def lookup_repo_storages_by_content_checksum(repo, checksums, by_manifest=False):
   """ Looks up repository storages (without placements) matching the given repository
       and checksum. """
   # There may be many duplicates of the checksums, so for performance reasons we are going
@@ -279,14 +279,29 @@ def lookup_repo_storages_by_content_checksum(repo, checksums):
 
   for counter, checksum in enumerate(set(checksums)):
     query_alias = 'q{0}'.format(counter)
-    candidate_subq = (ImageStorage
-                      .select(ImageStorage.id, ImageStorage.content_checksum,
-                              ImageStorage.image_size, ImageStorage.uuid, ImageStorage.cas_path,
-                              ImageStorage.uncompressed_size, ImageStorage.uploading)
-                      .join(Image)
-                      .where(Image.repository == repo, ImageStorage.content_checksum == checksum)
-                      .limit(1)
-                      .alias(query_alias))
+
+    # TODO(jschorr): Remove once we have a new-style model for tracking temp uploaded blobs and
+    # all legacy tables have been removed.
+    if by_manifest:
+      candidate_subq = (ImageStorage
+                        .select(ImageStorage.id, ImageStorage.content_checksum,
+                                ImageStorage.image_size, ImageStorage.uuid, ImageStorage.cas_path,
+                                ImageStorage.uncompressed_size, ImageStorage.uploading)
+                        .join(ManifestBlob)
+                        .where(ManifestBlob.repository == repo,
+                               ImageStorage.content_checksum == checksum)
+                        .limit(1)
+                        .alias(query_alias))
+    else:
+      candidate_subq = (ImageStorage
+                        .select(ImageStorage.id, ImageStorage.content_checksum,
+                                ImageStorage.image_size, ImageStorage.uuid, ImageStorage.cas_path,
+                                ImageStorage.uncompressed_size, ImageStorage.uploading)
+                        .join(Image)
+                        .where(Image.repository == repo, ImageStorage.content_checksum == checksum)
+                        .limit(1)
+                        .alias(query_alias))
+
     queries.append(ImageStorage
                    .select(SQL('*'))
                    .from_(candidate_subq))