Merge pull request crs4#152 from nfdi4plants/fix-cyclic-datasets

Fixed Check for link between root data entity an data entities

Author: kikkomep

rocrate_validator/profiles/ro-crate/must/4_data_entity_metadata.ttl

@@ -137,9 +137,12 @@ ro-crate:DataEntityRequiredPropertiesShape a sh:NodeShape ;
     sh:property
         [
             a sh:PropertyShape ;
-            sh:path [ sh:inversePath schema_org:hasPart ] ;
-            sh:node schema_org:Dataset ;
-            sh:minCount 1 ;
+            # sh:path [ sh:inversePath schema_org:hasPart ] ;
+            # sh:node schema_org:Dataset ;
+            # sh:minCount 1 ;
+            sh:path [ sh:oneOrMorePath [ sh:inversePath schema_org:hasPart ] ] ;
+            sh:qualifiedValueShape [ sh:class ro-crate:RootDataEntity ] ;
+            sh:qualifiedMinCount 1 ;
             sh:name "Data Entity MUST be directly referenced" ;
             sh:description """Check if the Data Entity is linked, either directly or indirectly, to the `Root Data Entity` using the `hasPart` (as defined in `schema.org`) property" """ ;
             # sh:message "A Data Entity MUST be directly or indirectly linked to the `Root Data Entity` through the `hasPart` property" ;

tests/data/crates/invalid/4_data_entity_metadata/cyclic_datasets/ro-crate-metadata.json

@@ -0,0 +1,69 @@
+{
+    "@context": [
+        "https://w3id.org/ro/crate/1.1/context"
+    ],
+    "@graph": [
+        {
+            "@id": "ro-crate-metadata.json",
+            "@type": "CreativeWork",
+            "conformsTo": {
+                "@id": "https://w3id.org/ro/crate/1.1"
+            },
+            "about": {
+                "@id": "./"
+            }
+        },
+        {
+            "@id": "./",
+            "@type": "Dataset",
+            "name": "Root",
+            "description": "Root Dataset",
+            "datePublished": "2024-05-17T01:04:52+01:00",
+            "hasPart": [
+            ],
+            "license": "CC-BY-4.0"
+        },
+        {
+            "@id": "dataset1/file1.txt",
+            "@type": "File",
+            "description": "A file",
+            "encodingFormat": "text/plain",
+            "name": "file1.txt"
+        },
+        {
+            "@id": "dataset2/file2.txt",
+            "@type": "File",
+            "description": "Another file",
+            "encodingFormat": "text/plain",
+            "name": "file2.txt"
+        },
+        {
+            "@id": "dataset1/",
+            "@type": "Dataset",
+            "name": "Dataset 1",
+            "description": "First dataset",
+            "hasPart": [
+                {
+                    "@id": "dataset1/file1.txt"
+                },
+                {
+                    "@id": "dataset2/"
+                }
+            ]
+        },
+        {
+            "@id": "dataset2/",
+            "@type": "Dataset",
+            "name": "Dataset 2",
+            "description": "Second dataset",
+            "hasPart": [
+                {
+                    "@id": "dataset2/file2.txt"
+                },
+                {
+                    "@id": "dataset1/"
+                }
+            ]
+        }
+    ]
+}

tests/integration/profiles/ro-crate/test_data_entity_metadata.py

@@ -48,6 +48,16 @@ def test_data_entity_must_be_directly_linked():
     )
 
 
+def test_data_entity_not_linked():
+    """Test a RO-Crate with datasets not linked to the root data entity."""
+    do_entity_test(
+        paths.dataset_not_linked_to_root,
+        models.Severity.REQUIRED,
+        False,
+        skip_checks=[SKIP_LOCAL_DATA_ENTITY_EXISTENCE_CHECK_IDENTIFIER]
+    )
+
+
 def test_data_entity_must_be_indirectly_linked():
     """Test a RO-Crate without a root data entity."""
     do_entity_test(

tests/ro_crates.py

@@ -246,6 +246,10 @@ def missing_hasPart_data_entity_reference(self) -> Path:
     def direct_hasPart_data_entity_reference(self) -> Path:
         return self.base_path / "valid_direct_hasPart_reference"
 
+    @property
+    def dataset_not_linked_to_root(self) -> Path:
+        return self.base_path / "cyclic_datasets"
+
     @property
     def indirect_hasPart_data_entity_reference(self) -> Path:
         return self.base_path / "valid_indirect_hasPart_reference"