refactor: persistence layer clean-up

DEV_ENV.md

@@ -51,11 +51,12 @@ The dev environment is initialized with AWS Secrets/S3 data in the [scripts/setu
 
 ### Make targets for running tests in dev
 
-| Command                      | Description                                   | Notes |
-| ---------------------------- | --------------------------------------------- | ----- |
-| `make local-unit-test`       | Run backend tests in the dev environment      |       |
-| `make local-functional-test` | Run functional tests in the dev environment   |       |
-| `make local-smoke-test`      | Run frontend/e2e tests in the dev environment |       |
+| Command                      | Description                                                                                  | Notes |
+|------------------------------|----------------------------------------------------------------------------------------------| ----- |
+| `make local-unit-test`       | Run backend tests in the local dockerized environment, against mock of persistence layer     |       |
+| `make local-integration-test`      | Run backend tests in the local dockerized environment, against dockerized database instance. |       |
+| `make local-functional-test` | Run functional tests in local dockerized environment                                         |       |
+| `make local-smoke-test`      | Run frontend/e2e tests in the local dockerized environment                                   |       |
 
 ### External dependencies
 

backend/layers/common/entities.py

@@ -3,6 +3,8 @@
 from enum import Enum
 from typing import List, Optional
 
+import uuid
+
 from dataclasses_json import dataclass_json
 
 
@@ -98,43 +100,34 @@ def empty():
 
 
 @dataclass
-class CollectionId:
+class EntityId:
     id: str
 
-    def __repr__(self) -> str:
-        return self.id
-
-
-@dataclass
-class CollectionVersionId:
-    id: str
+    def __init__(self, entity_id: str = None):
+        self.id = str(entity_id) if entity_id is not None else str(uuid.uuid4())
 
     def __repr__(self) -> str:
         return self.id
 
 
-@dataclass
-class DatasetId:
-    id: str
+class CollectionId(EntityId):
+    pass
 
-    def __repr__(self) -> str:
-        return self.id
 
+class CollectionVersionId(EntityId):
+    pass
 
-@dataclass
-class DatasetVersionId:
-    id: str
 
-    def __repr__(self) -> str:
-        return self.id
+class DatasetId(EntityId):
+    pass
 
 
-@dataclass
-class DatasetArtifactId:
-    id: str
+class DatasetVersionId(EntityId):
+    pass
 
-    def __repr__(self) -> str:
-        return self.id
+
+class DatasetArtifactId(EntityId):
+    pass
 
 
 @dataclass

backend/layers/persistence/orm.py

@@ -1,4 +1,4 @@
-from sqlalchemy import Column, DateTime, Enum, ForeignKey, String, Table
+from sqlalchemy import Column, DateTime, Enum, ForeignKey, String
 from sqlalchemy.dialects.postgresql import ARRAY, BOOLEAN, JSON, UUID
 from sqlalchemy.orm import registry
 from sqlalchemy.schema import MetaData
@@ -10,71 +10,61 @@
 
 
 @mapper_registry.mapped
-class Collection:
+class CollectionTable:
 
-    __table__ = Table(
-        "Collection",
-        mapper_registry.metadata,
-        Column("id", UUID(as_uuid=True), primary_key=True),
-        Column("version_id", UUID(as_uuid=True)),
-        Column("originally_published_at", DateTime),
-        Column("tombstoned", BOOLEAN),
-    )
+    __tablename__ = "Collection"
+
+    id = Column(UUID(as_uuid=True), primary_key=True)
+    version_id = Column(UUID(as_uuid=True))
+    originally_published_at = Column(DateTime)
+    tombstone = Column(BOOLEAN)
 
 
 @mapper_registry.mapped
-class CollectionVersion:
-
-    __table__ = Table(
-        "CollectionVersion",
-        mapper_registry.metadata,
-        Column("version_id", UUID(as_uuid=True), primary_key=True),
-        Column("collection_id", UUID(as_uuid=True)),
-        Column("metadata", JSON),
-        Column("owner", String),
-        Column("curator_name", String),
-        Column("publisher_metadata", JSON),
-        Column("published_at", DateTime),
-        Column("created_at", DateTime),
-        Column("datasets", ARRAY(UUID(as_uuid=True))),
-    )
+class CollectionVersionTable:
+
+    __tablename__ = "CollectionVersion"
+
+    id = Column(UUID(as_uuid=True), primary_key=True)
+    collection_id = Column(UUID(as_uuid=True))
+    collection_metadata = Column(JSON)
+    owner = Column(String)
+    curator_name = Column(String)
+    publisher_metadata = Column(JSON)
+    published_at = Column(DateTime)
+    created_at = Column(DateTime)
+    datasets = Column(ARRAY(UUID(as_uuid=True)))
 
 
 @mapper_registry.mapped
-class Dataset:
+class DatasetTable:
+
+    __tablename__ = "Dataset"
 
-    __table__ = Table(
-        "Dataset",
-        mapper_registry.metadata,
-        Column("dataset_id", UUID(as_uuid=True), primary_key=True),
-        Column("dataset_version_id", UUID(as_uuid=True)),
-        Column("published_at", DateTime),
-    )
+    id = Column(UUID(as_uuid=True), primary_key=True)
+    version_id = Column(UUID(as_uuid=True))
+    published_at = Column(DateTime)
 
 
 @mapper_registry.mapped
-class DatasetVersion:
+class DatasetVersionTable:
 
-    __table__ = Table(
-        "DatasetVersion",
-        mapper_registry.metadata,
-        Column("version_id", UUID(as_uuid=True), primary_key=True),
-        Column("dataset_id", UUID(as_uuid=True), ForeignKey("Dataset.dataset_id")),
-        Column("collection_id", UUID(as_uuid=True)),
-        Column("created_at", DateTime),
-        Column("metadata", JSON),
-        Column("artifacts", ARRAY(UUID(as_uuid=True))),
-        Column("status", JSON),
-    )
+    __tablename__ = "DatasetVersion"
+
+    id = Column(UUID(as_uuid=True), primary_key=True)
+    dataset_id = Column(UUID(as_uuid=True), ForeignKey("Dataset.id"))
+    collection_id = Column(UUID(as_uuid=True))
+    created_at = Column(DateTime)
+    dataset_metadata = Column(JSON)
+    artifacts = Column(ARRAY(UUID(as_uuid=True)))
+    status = Column(JSON)
 
 
 @mapper_registry.mapped
-class DatasetArtifact:
-
-    __table__ = Table(
-        "DatasetArtifact",
-        mapper_registry.metadata,
-        Column("id", UUID(as_uuid=True), primary_key=True),
-        Column("type", Enum(DatasetArtifactType)),
-        Column("uri", String),
-    )
+class DatasetArtifactTable:
+
+    __tablename__ = "DatasetArtifact"
+
+    id = Column(UUID(as_uuid=True), primary_key=True)
+    type = Column(Enum(DatasetArtifactType))
+    uri = Column(String)