Fix a bug that appeared when calling ".exists()" on a ModelVersion (which did exist in the KG) with a repository that had no @id.

The fix requires openminds >= 0.5.1

The error trace looked like:
```
  /Users/adavison/dev/data/fairgraph/fairgraph/kgobject.py(566)exists()
-> self._update_empty_properties(data)  # also updates `remote_data`
  /Users/adavison/dev/data/fairgraph/fairgraph/kgobject.py(502)_update_empty_properties()
-> locally_modified = self.modified_data()
> /Users/adavison/dev/data/fairgraph/fairgraph/kgobject.py(650)modified_data()
-> self.to_jsonld(include_empty_properties=True, embed_linked_nodes=False), self.context
  /Users/adavison/dev/validation/env/lib/python3.9/site-packages/openminds/base.py(107)to_jsonld()
-> data[property.path] = value_to_jsonld(
  /Users/adavison/dev/validation/env/lib/python3.9/site-packages/openminds/base.py(30)value_to_jsonld()
-> raise ValueError("Exporting as a stand-alone JSON-LD document requires @id to be defined.")
```

Author: apdavison

fairgraph/embedded.py

@@ -23,6 +23,8 @@
 from typing import Optional, TYPE_CHECKING, Union
 from warnings import warn
 
+from openminds.base import LinkedNodeEmbedding
+
 from .utility import as_list, ActivityLog
 from .base import JSONdict
 from .node import KGNode
@@ -66,8 +68,8 @@ def __repr__(self):
         return template.format(self=self)
 
     def __eq__(self, other):
-        return isinstance(other, self.__class__) and self.to_jsonld(embed_linked_nodes=False) == other.to_jsonld(
-            embed_linked_nodes=False
+        return isinstance(other, self.__class__) and self.to_jsonld(embed_linked_nodes=LinkedNodeEmbedding.NEVER) == other.to_jsonld(
+            embed_linked_nodes=LinkedNodeEmbedding.NEVER
         )
 
     @classmethod

fairgraph/kgobject.py

@@ -36,6 +36,7 @@
 
 from openminds.registry import lookup_type
 from openminds import IRI, LinkedMetadata
+from openminds.base import LinkedNodeEmbedding
 
 from .utility import expand_uri, as_list, expand_filter, ActivityLog, normalize_data, handle_scope_keyword
 from .queries import Query, QueryProperty
@@ -96,7 +97,7 @@ def __init__(
             if data:
                 try:
                     self.remote_data = normalize_data(
-                        self.to_jsonld(include_empty_properties=False, embed_linked_nodes=False),
+                        self.to_jsonld(include_empty_properties=False, embed_linked_nodes=LinkedNodeEmbedding.NEVER),
                         data.get("@context", self.context)
                     )
                 except ValueError as err:
@@ -647,7 +648,11 @@ def values_are_equal(local, remote):
                 return local == remote
 
         current_data = normalize_data(
-            self.to_jsonld(include_empty_properties=True, embed_linked_nodes=False), self.context
+            self.to_jsonld(
+                include_empty_properties=True,
+                embed_linked_nodes=LinkedNodeEmbedding.IF_NECESSARY
+            ),
+            self.context
         )
         modified_data = {}
         for key, current_value in current_data.items():
@@ -741,14 +746,15 @@ def save(
             else:
                 # update
                 local_data = normalize_data(
-                    self.to_jsonld(include_empty_properties=False, embed_linked_nodes=False),
+                    self.to_jsonld(include_empty_properties=False, embed_linked_nodes=LinkedNodeEmbedding.NEVER),
                     self.context
                 )
                 if replace:
                     logger.info(f"  - replacing - {self.__class__.__name__}(id={self.id})")
                     if activity_log:
                         activity_log.update(item=self, delta=local_data, space=space, entry_type="replacement")
                     try:
+                        assert self.uuid is not None
                         client.replace_instance(self.uuid, local_data)
                         # what does this return? Can we use it to update `remote_data`?
                     except AuthorizationError as err:
@@ -778,6 +784,7 @@ def save(
                                 # Note: if modified_data includes embedded objects
                                 # then _all_ fields of the embedded objects must be provided,
                                 # not only those that have changed.
+                                assert self.uuid is not None
                                 client.update_instance(self.uuid, modified_data)
                             except AuthorizationError as err:
                                 if ignore_auth_errors:
@@ -800,7 +807,7 @@ def save(
         else:
             # create new
             local_data = normalize_data(
-                self.to_jsonld(include_empty_properties=False, embed_linked_nodes=False),
+                self.to_jsonld(include_empty_properties=False, embed_linked_nodes=LinkedNodeEmbedding.NEVER),
                 self.context
             )
             logger.info("  - creating instance with data {}".format(local_data))
@@ -836,7 +843,7 @@ def save(
         if self.id:
             logger.debug(
                 "Updating cache for object {}. Current state: {}".format(
-                    self.id, self.to_jsonld(embed_linked_nodes=False)
+                    self.id, self.to_jsonld(embed_linked_nodes=LinkedNodeEmbedding.NEVER)
                 )
             )
             object_cache[self.id] = self

fairgraph/node.py

@@ -5,7 +5,7 @@
 import logging
 from warnings import warn
 
-from openminds.base import value_to_jsonld, Link
+from openminds.base import value_to_jsonld, Link, LinkedNodeEmbedding
 from openminds.properties import Property
 
 from .registry import NodeMeta
@@ -480,7 +480,7 @@ def _build_existence_query(self) -> Union[None, Dict[str, Any]]:
             elif value is None:
                 raise CannotBuildExistenceQuery(f"Required value for '{query_property_name}' is missing")
             else:
-                query_val = value_to_jsonld(value, include_empty_properties=False, embed_linked_nodes=False)
+                query_val = value_to_jsonld(value, include_empty_properties=False, embed_linked_nodes=LinkedNodeEmbedding.NEVER)
                 if query_val is None:
                     raise CannotBuildExistenceQuery(f"Required value for '{query_property_name}' is missing")
                 query[query_property_name] = query_val

pyproject.toml

@@ -17,7 +17,7 @@ classifiers = [
 keywords = ["ebrains", "hbp", "metadata", "electrophysiology", "knowledge-graph"]
 dependencies=[
     "ebrains-kg-core",
-    "openminds>=0.4.0",
+    "openminds>=0.5.1",
     "python-dateutil",
     "tabulate",
     "requests"

test/test_base.py

@@ -5,7 +5,8 @@
 
 from datetime import date, datetime
 from numbers import Real
-from openminds.base import LinkedMetadata, EmbeddedMetadata as OMEmbeddedMetadata
+from openminds.base import LinkedMetadata, EmbeddedMetadata as OMEmbeddedMetadata, LinkedNodeEmbedding
+
 from openminds.properties import Property
 from fairgraph.embedded import KGEmbedded
 from fairgraph.kgobject import KGObject
@@ -473,7 +474,7 @@ def test_build_data_all_properties(self):
             "https://openminds.ebrains.eu/vocab/anOptionalListOfStrings": ["plum, peach, apricot"],
             "https://openminds.ebrains.eu/vocab/anOptionalString": "melon",
         }
-        assert obj.to_jsonld(include_empty_properties=False, embed_linked_nodes=False) == expected
+        assert obj.to_jsonld(include_empty_properties=False, embed_linked_nodes=LinkedNodeEmbedding.NEVER) == expected
 
     def test_modified_data(self):
         obj = self._construct_object_all_properties()
@@ -527,7 +528,7 @@ def test_exists__it_does_exist(self):
 
         class MockClient:
             def instance_from_full_uri(self, id, use_cache=True, release_status="in progress", require_full_data=True):
-                data = orig_object.to_jsonld(include_empty_properties=False, embed_linked_nodes=False)
+                data = orig_object.to_jsonld(include_empty_properties=False, embed_linked_nodes=LinkedNodeEmbedding.NEVER)
                 data["https://core.kg.ebrains.eu/vocab/meta/space"] = "collab-foobar"
                 data["@id"] = orig_object.id
                 data["@type"] = orig_object.type_

test/test_openminds_core.py

@@ -11,14 +11,15 @@
 import pytest
 
 from openminds import IRI
+from openminds.base import LinkedNodeEmbedding
 
 from fairgraph.utility import as_list
 from fairgraph.kgproxy import KGProxy
 from fairgraph.kgquery import KGQuery
 from fairgraph.kgobject import KGObject
 import fairgraph.openminds.core as omcore
 import fairgraph.openminds.controlled_terms as omterms
-from fairgraph.utility import ActivityLog, sha1sum
+from fairgraph.utility import ActivityLog, sha1sum, normalize_data
 
 from test.utils import mock_client, kg_client, skip_if_no_connection, skip_if_using_production_server
 
@@ -433,6 +434,40 @@ def test_exists_method_without_id(kg_client):
     assert new_model == model
 
 
+def test_modified_data_method_with_local_changes():
+    dsv = omcore.DatasetVersion(
+        id="https://kg.ebrains.eu/api/instances/12345678-90ab-cdef-0123-4567890abcde", full_name="test1234"
+    )
+    dsv.remote_data = normalize_data(
+        dsv.to_jsonld(include_empty_properties=True, embed_linked_nodes=LinkedNodeEmbedding.NEVER), dsv.context
+    )
+    assert dsv.modified_data() == {}
+    dsv.repository = omcore.FileRepository(iri="http://example.org")
+    assert dsv.modified_data() == {
+        "https://openminds.om-i.org/props/repository": {
+            "@type": "https://openminds.om-i.org/types/FileRepository",
+            "https://openminds.om-i.org/props/IRI": "http://example.org",
+            "https://openminds.om-i.org/props/contentTypePattern": None,
+            "https://openminds.om-i.org/props/format": None,
+            "https://openminds.om-i.org/props/hash": None,
+            "https://openminds.om-i.org/props/hostedBy": None,
+            "https://openminds.om-i.org/props/name": None,
+            "https://openminds.om-i.org/props/storageSize": None,
+            "https://openminds.om-i.org/props/structurePattern": None,
+            "https://openminds.om-i.org/props/type": None,
+        }
+    }
+    dsv.repository = omcore.FileRepository(
+        id="https://kg.ebrains.eu/api/instances/23456789-0abc-def0-1234-567890abcdef",
+        iri="http://example.org"
+    )
+    assert dsv.modified_data() == {
+        "https://openminds.om-i.org/props/repository": {
+            "@id": dsv.repository.id
+        }
+    }
+
+
 def test__update():
     example_data = {
         "@id": "https://kg.ebrains.eu/api/instances/e90fc25a-fc35-4066-9ff2-ca3583a2d008",
@@ -518,9 +553,11 @@ def test_to_jsonld():
         "familyName": "Oakenshield",
         "givenName": "Thorin",
     }
-    assert person1.to_jsonld(embed_linked_nodes=True, include_empty_properties=False) == expected1
+    assert (
+        person1.to_jsonld(embed_linked_nodes=LinkedNodeEmbedding.ALWAYS, include_empty_properties=False) == expected1
+    )
     with pytest.raises(ValueError) as err:
-        person1.to_jsonld(embed_linked_nodes=False, include_empty_properties=False)
+        person1.to_jsonld(embed_linked_nodes=LinkedNodeEmbedding.NEVER, include_empty_properties=False)
         assert "Exporting as a stand-alone JSON-LD document requires @id to be defined" in err
 
     person2 = omcore.Person(
@@ -559,8 +596,12 @@ def test_to_jsonld():
         "familyName": "Oakenshield",
         "givenName": "Thorin",
     }
-    assert person2.to_jsonld(embed_linked_nodes=True, include_empty_properties=False) == expected2a
-    assert person2.to_jsonld(embed_linked_nodes=False, include_empty_properties=False) == expected2b
+    assert (
+        person2.to_jsonld(embed_linked_nodes=LinkedNodeEmbedding.ALWAYS, include_empty_properties=False) == expected2a
+    )
+    assert (
+        person2.to_jsonld(embed_linked_nodes=LinkedNodeEmbedding.NEVER, include_empty_properties=False) == expected2b
+    )
 
 
 @skip_if_using_production_server