Promote resource to a top-level entity

migrate/versions/2023_11_20_220060b4adb3_create_archive_and_package_entities.py → migrate/versions/2023_12_04_002ebdb598a3_create_archive_resource_and_package_.py

@@ -1,16 +1,16 @@
-"""Create archive and package entities
+"""Create archive, resource and package entities
 
-Revision ID: 220060b4adb3
+Revision ID: 002ebdb598a3
 Revises: df57d06e1ee5
-Create Date: 2023-11-20 14:59:44.964701
+Create Date: 2023-12-04 09:21:03.086156
 
 """
 from alembic import op
 import sqlalchemy as sa
 from sqlalchemy.dialects import postgresql
 
 # revision identifiers, used by Alembic.
-revision = '220060b4adb3'
+revision = '002ebdb598a3'
 down_revision = 'df57d06e1ee5'
 branch_labels = None
 depends_on = None
@@ -23,42 +23,43 @@ def upgrade():
                     sa.Column('url', sa.String(), nullable=False),
                     sa.PrimaryKeyConstraint('id')
                     )
-    op.create_table('archive_resource',
+    op.create_table('resource',
                     sa.Column('id', sa.String(), nullable=False),
-                    sa.Column('key', sa.String(), nullable=False),
-                    sa.Column('path', sa.String(), nullable=True),
+                    sa.Column('type', sa.Enum('file', 'folder', 'webpage', name='resourcetype'), nullable=False),
+                    sa.Column('path', sa.String(), nullable=False),
                     sa.Column('name', sa.String(), nullable=True),
-                    sa.Column('type', sa.String(), nullable=True),
-                    sa.Column('size', sa.Integer(), nullable=True),
+                    sa.Column('size', sa.BigInteger(), nullable=True),
                     sa.Column('md5', sa.String(), nullable=True),
                     sa.Column('timestamp', sa.TIMESTAMP(timezone=True), nullable=True),
                     sa.Column('archive_id', sa.String(), nullable=False),
                     sa.ForeignKeyConstraint(['archive_id'], ['archive.id'], ondelete='CASCADE'),
-                    sa.PrimaryKeyConstraint('id'),
-                    sa.UniqueConstraint('archive_id', 'key', name='uix_archive_resource_archive_id_key')
+                    sa.PrimaryKeyConstraint('id')
                     )
     op.create_table('package',
                     sa.Column('id', sa.String(), nullable=False),
                     sa.Column('metadata_', postgresql.JSONB(astext_type=sa.Text()), nullable=True),
                     sa.Column('timestamp', sa.TIMESTAMP(timezone=True), nullable=False),
+                    sa.Column('provider_id', sa.String(), nullable=False),
                     sa.Column('record_id', sa.String(), nullable=True),
+                    sa.ForeignKeyConstraint(['provider_id'], ['provider.id'], ondelete='CASCADE'),
                     sa.ForeignKeyConstraint(['record_id'], ['record.id'], ondelete='SET NULL'),
                     sa.PrimaryKeyConstraint('id')
                     )
     op.create_table('package_resource',
                     sa.Column('package_id', sa.String(), nullable=False),
                     sa.Column('resource_id', sa.String(), nullable=False),
                     sa.ForeignKeyConstraint(['package_id'], ['package.id'], ondelete='CASCADE'),
-                    sa.ForeignKeyConstraint(['resource_id'], ['archive_resource.id'], ondelete='RESTRICT'),
+                    sa.ForeignKeyConstraint(['resource_id'], ['resource.id'], ondelete='RESTRICT'),
                     sa.PrimaryKeyConstraint('package_id', 'resource_id')
                     )
     # ### end Alembic commands ###
 
 
 def downgrade():
-    # ### commands auto generated by Alembic ###
+    # ### commands auto generated by Alembic - added drop type ###
     op.drop_table('package_resource')
     op.drop_table('package')
-    op.drop_table('archive_resource')
+    op.drop_table('resource')
     op.drop_table('archive')
+    op.execute('drop type resourcetype')
     # ### end Alembic commands ###

odp/db/models/__init__.py

@@ -1,14 +1,15 @@
-from .archive import Archive, ArchiveResource
+from .archive import Archive
 from .catalog import Catalog, CatalogRecord, CatalogRecordFacet
 from .client import Client, ClientCollection, ClientScope
 from .collection import Collection, CollectionAudit, CollectionTag, CollectionTagAudit
 from .package import Package, PackageResource
 from .provider import Provider, ProviderAudit
 from .record import PublishedRecord, Record, RecordAudit, RecordTag, RecordTagAudit
+from .resource import Resource
 from .role import Role, RoleCollection, RoleScope
 from .schema import Schema
 from .scope import Scope
 from .tag import Tag
-from .types import AuditCommand, IdentityCommand, SchemaType, ScopeType, TagCardinality, TagType
+from .types import AuditCommand, IdentityCommand, ResourceType, SchemaType, ScopeType, TagCardinality, TagType
 from .user import IdentityAudit, User, UserRole
 from .vocabulary import Vocabulary, VocabularyTerm, VocabularyTermAudit

odp/db/models/archive.py

@@ -1,7 +1,4 @@
-import uuid
-
-from sqlalchemy import BigInteger, Column, ForeignKey, String, TIMESTAMP, UniqueConstraint
-from sqlalchemy.orm import relationship
+from sqlalchemy import Column, String
 
 from odp.db import Base
 
@@ -16,34 +13,3 @@ class Archive(Base):
     url = Column(String, nullable=False)
 
     _repr_ = 'id', 'url'
-
-
-class ArchiveResource(Base):
-    """A reference to a file or dataset in an archive.
-
-    `key` is the archive's unique identifier for the resource.
-    `path`, if specified, is relative to `url` of the archive.
-    """
-
-    __tablename__ = 'archive_resource'
-
-    __table_args__ = (
-        UniqueConstraint(
-            'archive_id', 'key',
-            name='uix_archive_resource_archive_id_key',
-        ),
-    )
-
-    id = Column(String, primary_key=True, default=lambda: str(uuid.uuid4()))
-    key = Column(String, nullable=False)
-    path = Column(String)
-    name = Column(String)
-    type = Column(String)
-    size = Column(BigInteger)
-    md5 = Column(String)
-    timestamp = Column(TIMESTAMP(timezone=True))
-
-    archive_id = Column(String, ForeignKey('archive.id', ondelete='CASCADE'), nullable=False)
-    archive = relationship('Archive')
-
-    _repr_ = 'id', 'key', 'path', 'name', 'type', 'size', 'archive_id'

odp/db/models/package.py

@@ -32,12 +32,15 @@ class Package(Base):
 
 class PackageResource(Base):
     """Model of a many-to-many package-resource association,
-    representing the set of archived resources constituting a package."""
+    representing the set of resources constituting a package.
+
+    A resource cannot be deleted if it is part of a package.
+    """
 
     __tablename__ = 'package_resource'
 
     package_id = Column(String, ForeignKey('package.id', ondelete='CASCADE'), primary_key=True)
-    resource_id = Column(String, ForeignKey('archive_resource.id', ondelete='RESTRICT'), primary_key=True)
+    resource_id = Column(String, ForeignKey('resource.id', ondelete='RESTRICT'), primary_key=True)
 
     package = relationship('Package', viewonly=True)
-    resource = relationship('ArchiveResource', viewonly=True)
+    resource = relationship('Resource', viewonly=True)

odp/db/models/resource.py

@@ -0,0 +1,29 @@
+import uuid
+
+from sqlalchemy import BigInteger, Column, Enum, ForeignKey, String, TIMESTAMP
+from sqlalchemy.orm import relationship
+
+from odp.db import Base
+from odp.db.models.types import ResourceType
+
+
+class Resource(Base):
+    """A reference to a file, folder or web page in an archive.
+
+    `path` is relative to `url` of the archive.
+    """
+
+    __tablename__ = 'resource'
+
+    id = Column(String, primary_key=True, default=lambda: str(uuid.uuid4()))
+    type = Column(Enum(ResourceType), nullable=False)
+    path = Column(String, nullable=False)
+    name = Column(String)
+    size = Column(BigInteger)
+    md5 = Column(String)
+    timestamp = Column(TIMESTAMP(timezone=True))
+
+    archive_id = Column(String, ForeignKey('archive.id', ondelete='CASCADE'), nullable=False)
+    archive = relationship('Archive')
+
+    _repr_ = 'id', 'path', 'name', 'type', 'size', 'archive_id'

odp/db/models/types.py

@@ -6,6 +6,12 @@ def __repr__(self):
         return repr(self.value)
 
 
+class ResourceType(DBEnum):
+    file = 'file'
+    folder = 'folder'
+    webpage = 'webpage'
+
+
 class SchemaType(DBEnum):
     metadata = 'metadata'
     tag = 'tag'