Remove descriptor and descriptor_schema from Entity

docs/CHANGELOG.rst

@@ -13,6 +13,8 @@ Unreleased
 Changed
 -------
 - **BACKWARD INCOMPATIBLE:** Make move between collections a background job
+- **BACKWARD INCOMPATIBLE:** Remove ``descriptor_schema`` and ``descriptor``
+  field from the ``Entity`` model
 - Require ``Jinja >= 3.1.3`` to address security vulnerability
   https://deps.dev/advisory/osv/GHSA-h5c8-rqwp-cp95
 - Do not add ``descriptor_schema`` to automatically created entities

docs/example/example/processes/template_py_process.py

@@ -62,11 +62,9 @@ class YourProcessName(Process):
     # type sample. See
     # https://resolwe.readthedocs.io/en/latest/proc.html#entity for
     # more details.
-    # Fields descriptor_schema, input and always_create are optional.
+    # Field always_create is optional.
     entity = {
         "type": "sample",  # name of your entity
-        # Use a descriptor schema slug if it differs from "type".
-        "descriptor_schema": "descriptor-schema-of-entity",
         "input": "name of input that has the entity",
         # If you want to create entity in all cases
         "always_create": True,

resolwe/flow/filters.py

@@ -303,7 +303,6 @@ class Meta(BaseResolweFilter.Meta):
             **BaseResolweFilter.Meta.fields,
             **{
                 "description": TEXT_LOOKUPS[:],
-                "descriptor_schema": ["exact"],
             },
         }
 
@@ -331,6 +330,12 @@ class Meta(BaseCollectionFilter.Meta):
         """Filter configuration."""
 
         model = Collection
+        fields = {
+            **BaseCollectionFilter.Meta.fields,
+            **{
+                "descriptor_schema": ["exact"],
+            },
+        }
 
     def count_entities(self, queryset: QuerySet, name: str, value: str):
         """Filter by the number of associated entities."""

resolwe/flow/management/commands/register.py

@@ -184,22 +184,10 @@ def register_processes(self, process_schemas, user, force=False, verbosity=1):
                     continue
 
                 p["entity_type"] = p["entity"]["type"]
-                p["entity_descriptor_schema"] = p["entity"].get(
-                    "descriptor_schema", p["entity_type"]
-                )
                 p["entity_input"] = p["entity"].get("input", None)
                 p["entity_always_create"] = p["entity"].get("always_create", False)
                 p.pop("entity")
 
-                if not DescriptorSchema.objects.filter(
-                    slug=p["entity_descriptor_schema"]
-                ).exists():
-                    self.stderr.write(
-                        "Skip processor {}: Unknown descriptor schema '{}' used in 'entity' "
-                        "field.".format(p["slug"], p["entity_descriptor_schema"])
-                    )
-                    continue
-
             if "persistence" in p:
                 persistence_mapping = {
                     "RAW": Process.PERSISTENCE_RAW,
@@ -452,8 +440,6 @@ def handle(self, *args, **options):
         self.register_descriptors(
             descriptor_schemas, user_admin, force, verbosity=verbosity
         )
-        # NOTE: Descriptor schemas must be registered first, so
-        #       processes can validate 'entity_descriptor_schema' field.
         self.register_processes(process_schemas, user_admin, force, verbosity=verbosity)
 
         if retire:

resolwe/flow/migrations/0023_remove_entity_descriptor_and_more.py

@@ -0,0 +1,29 @@
+# Generated by Django 4.2.11 on 2024-03-18 13:12
+
+from django.db import migrations
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("flow", "0022_fix_group_by"),
+    ]
+
+    operations = [
+        migrations.RemoveField(
+            model_name="entity",
+            name="descriptor",
+        ),
+        migrations.RemoveField(
+            model_name="entity",
+            name="descriptor_dirty",
+        ),
+        migrations.RemoveField(
+            model_name="entity",
+            name="descriptor_schema",
+        ),
+        migrations.RemoveField(
+            model_name="process",
+            name="entity_descriptor_schema",
+        ),
+    ]

resolwe/flow/migrations/triggers_entity.sql

@@ -26,8 +26,6 @@ CREATE OR REPLACE FUNCTION generate_resolwe_entity_search(entity flow_entity)
         FROM auth_user
         WHERE id = entity.contributor_id;
 
-        SELECT COALESCE(flatten_descriptor_values(entity.descriptor), '') INTO flat_descriptor;
-
         SELECT
             -- Entity name.
             setweight(to_tsvector('simple', entity.name), 'A') ||
@@ -52,12 +50,7 @@ CREATE OR REPLACE FUNCTION generate_resolwe_entity_search(entity flow_entity)
             -- Owners last names. There is no guarantee that it is not NULL.
             setweight(to_tsvector('simple', COALESCE(owners.last_names, '')), 'B') ||
             -- Entity tags.
-            setweight(to_tsvector('simple', array_to_string(entity.tags, ' ')), 'B') ||
-            -- Entity descriptor.
-            setweight(to_tsvector('simple', flat_descriptor), 'C') ||
-            setweight(to_tsvector('simple', get_characters(flat_descriptor)), 'D') ||
-            setweight(to_tsvector('simple', get_numbers(flat_descriptor)), 'D')
-
+            setweight(to_tsvector('simple', array_to_string(entity.tags, ' ')), 'B')
         INTO search;
 
         RETURN search;

resolwe/flow/models/collection.py

@@ -30,37 +30,12 @@ class Meta(BaseModel.Meta):
 
     settings = models.JSONField(default=dict)
 
-    #: collection descriptor schema
-    descriptor_schema = models.ForeignKey(
-        "flow.DescriptorSchema", blank=True, null=True, on_delete=models.PROTECT
-    )
-
-    #: collection descriptor
-    descriptor = models.JSONField(default=dict)
-
-    #: indicate whether `descriptor` doesn't match `descriptor_schema` (is dirty)
-    descriptor_dirty = models.BooleanField(default=False)
-
     #: tags for categorizing objects
     tags = ArrayField(models.CharField(max_length=255), default=list)
 
     #: field used for full-text search
     search = SearchVectorField(null=True)
 
-    def save(self, *args, **kwargs):
-        """Perform descriptor validation and save object."""
-        if self.descriptor_schema:
-            try:
-                validate_schema(self.descriptor, self.descriptor_schema.schema)
-                self.descriptor_dirty = False
-            except DirtyError:
-                self.descriptor_dirty = True
-        elif self.descriptor and self.descriptor != {}:
-            raise ValueError(
-                "`descriptor_schema` must be defined if `descriptor` is given"
-            )
-        super().save(*args, **kwargs)
-
 
 class CollectionQuerySet(BaseQuerySet, PermissionQuerySet):
     """Query set for ``Collection`` objects."""
@@ -122,6 +97,17 @@ class Meta(BaseCollection.Meta):
         AnnotationField, related_name="collection"
     )
 
+    #: collection descriptor schema
+    descriptor_schema = models.ForeignKey(
+        "flow.DescriptorSchema", blank=True, null=True, on_delete=models.PROTECT
+    )
+
+    #: collection descriptor
+    descriptor = models.JSONField(default=dict)
+
+    #: indicate whether `descriptor` doesn't match `descriptor_schema` (is dirty)
+    descriptor_dirty = models.BooleanField(default=False)
+
     def is_duplicate(self):
         """Return True if collection is a duplicate."""
         return bool(self.duplicated)
@@ -147,8 +133,18 @@ def delete_background(self):
         )
 
     def save(self, *args, **kwargs):
-        """Add required annotation fields to the collection."""
+        """Perform descriptor validation and save object."""
         create = self.pk is None
+        if self.descriptor_schema:
+            try:
+                validate_schema(self.descriptor, self.descriptor_schema.schema)
+                self.descriptor_dirty = False
+            except DirtyError:
+                self.descriptor_dirty = True
+        elif self.descriptor and self.descriptor != {}:
+            raise ValueError(
+                "`descriptor_schema` must be defined if `descriptor` is given"
+            )
         super().save(*args, **kwargs)
         if create:
             required_fields = AnnotationField.objects.filter(required=True)

resolwe/flow/models/data.py

@@ -32,7 +32,6 @@
 from resolwe.permissions.utils import assign_contributor_permissions, copy_permissions
 
 from .base import BaseModel, BaseQuerySet
-from .descriptor import DescriptorSchema
 from .entity import Entity, EntityQuerySet
 from .history_manager import HistoryMixin
 from .secret import Secret

resolwe/flow/models/process.py

@@ -156,12 +156,6 @@ class Meta(BaseModel.Meta):
     create a new one.
     """
 
-    entity_descriptor_schema = models.CharField(max_length=100, null=True, blank=True)
-    """
-    Slug of the descriptor schema assigned to the Entity created with
-    :attr:`~resolwe.flow.models.Process.entity_type`.
-    """
-
     entity_input = models.CharField(max_length=100, null=True, blank=True)
     """
     Limit the entity selection in
@@ -170,8 +164,7 @@ class Meta(BaseModel.Meta):
 
     entity_always_create = models.BooleanField(default=False)
     """
-    Create new entity, regardless of ``entity_input`` or
-    ``entity_descriptor_schema`` fields.
+    Create new entity, regardless of ``entity_input`` field.
     """
 
     run = models.JSONField(default=dict)

resolwe/flow/serializers/collection.py

@@ -76,7 +76,6 @@ class Meta:
         model = Collection
         read_only_fields = (
             "created",
-            "descriptor_dirty",
             "duplicated",
             "id",
             "modified",
@@ -130,4 +129,5 @@ class Meta(BaseCollectionSerializer.Meta):
             "entity_count",
             "descriptor",
             "descriptor_schema",
+            "descriptor_dirty",
         )

resolwe/flow/serializers/process.py

@@ -24,7 +24,6 @@ class Meta:
             "data_name",
             "description",
             "entity_always_create",
-            "entity_descriptor_schema",
             "entity_input",
             "entity_type",
             "input_schema",

resolwe/flow/tests/test_api.py

@@ -53,7 +53,6 @@ def setUp(self):
             slug="test-process",
             version="1.0.0",
             contributor=self.contributor,
-            entity_descriptor_schema="test-schema",
             input_schema=[
                 {"name": "input_data", "type": "data:test:", "required": False}
             ],
@@ -92,10 +91,7 @@ def setUp(self):
         )
 
         self.entity = Entity.objects.create(
-            collection=self.collection,
-            contributor=self.contributor,
-            descriptor_schema=self.descriptor_schema,
-            name="Test entity",
+            collection=self.collection, contributor=self.contributor, name="Test entity"
         )
         self.data = Data.objects.create(
             name="Test data",
@@ -380,7 +376,6 @@ def setUp(self):
             version="1.0.0",
             contributor=self.contributor,
             entity_type="test-schema",
-            entity_descriptor_schema="test-schema",
             input_schema=[
                 {"name": "input_data", "type": "data:test:", "required": False}
             ],
@@ -395,9 +390,7 @@ def setUp(self):
             descriptor_schema=self.descriptor_schema,
         )
         self.entity = Entity.objects.create(
-            collection=self.collection,
-            contributor=self.contributor,
-            descriptor_schema=self.descriptor_schema,
+            collection=self.collection, contributor=self.contributor
         )
         self.collection.set_permission(Permission.EDIT, self.contributor)
         self.proc.set_permission(Permission.VIEW, self.contributor)
@@ -565,11 +558,7 @@ def test_prefetch(self):
         collection_2 = Collection.objects.create(
             contributor=self.user, descriptor_schema=descriptor_schema_2
         )
-        entity_2 = Entity.objects.create(
-            collection=collection_2,
-            contributor=self.user,
-            descriptor_schema=descriptor_schema_2,
-        )
+        entity_2 = Entity.objects.create(collection=collection_2, contributor=self.user)
 
         for i in range(5):
             create_kwargs = {
@@ -607,7 +596,7 @@ def test_prefetch(self):
         with CaptureQueriesContext(conn) as captured_queries:
             response = self.data_viewset(request)
             self.assertEqual(len(response.data), 10)
-            self.assertEqual(len(captured_queries), 12)
+            self.assertIn(len(captured_queries), [12, 13])
 
     def test_descriptor_schema(self):
         # Descriptor schema can be assigned by slug.
@@ -1647,10 +1636,7 @@ def test_prefetch(self):
         )
 
         for i in range(5):
-            create_kwargs = {
-                "contributor": self.contributor,
-                "descriptor_schema": self.descriptor_schema,
-            }
+            create_kwargs = {"contributor": self.contributor}
             if i < 4:
                 create_kwargs["collection"] = self.collection
             entity = Entity.objects.create(**create_kwargs)
@@ -1659,10 +1645,7 @@ def test_prefetch(self):
             )
 
         for i in range(5):
-            create_kwargs = {
-                "contributor": self.user,
-                "descriptor_schema": descriptor_schema_2,
-            }
+            create_kwargs = {"contributor": self.user}
             if i < 4:
                 create_kwargs["collection"] = collection_2
             entity = Entity.objects.create(**create_kwargs)
@@ -1684,7 +1667,7 @@ def test_prefetch(self):
         with CaptureQueriesContext(conn) as captured_queries:
             response = self.entity_list_viewset(request)
             self.assertEqual(len(response.data), 10)
-            self.assertEqual(len(captured_queries), 7)
+            self.assertIn(len(captured_queries), [7, 8])
 
     def test_list_filter_collection(self):
         request = factory.get("/", {}, format="json")