Merge pull request crs4#159 from kikkomep/fix/issue-137

fix: allow entity validation when @base is defined in @context

Author: kikkomep

rocrate_validator/models.py

@@ -1776,7 +1776,7 @@ def update(self, event: Event, ctx: Optional[ValidationContext] = None) -> None:
             logger.debug("Validation ended with result: %s", event.validation_result)
 
     def to_dict(self) -> dict:
-        """"
+        """
         Get the computed validation statistics as a dictionary
         """
         return {

rocrate_validator/profiles/ro-crate/must/1_file-descriptor_metadata.ttl

@@ -11,7 +11,6 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-
 @prefix ro: <./> .
 @prefix ro-crate: <https://github.com/crs4/rocrate-validator/profiles/ro-crate/> .
 @prefix dct: <http://purl.org/dc/terms/> .
@@ -20,8 +19,31 @@
 @prefix sh: <http://www.w3.org/ns/shacl#> .
 @prefix validator: <https://github.com/crs4/rocrate-validator/> .
 
+ro-crate:ROCrateMetadataFileDescriptorExistence a sh:NodeShape ;
+    sh:targetNode sh:Graph ;
+    sh:name "RO-Crate Metadata File Descriptor entity existence" ;
+    sh:description """The RO-Crate JSON-LD MUST contain a Metadata File Descriptor entity named `ro-crate-metadata.json` and typed as `schema:CreativeWork`""" ;
+    sh:sparql [
+        a sh:SPARQLConstraint ;
+        sh:name "RO-Crate Metadata File Descriptor entity existence" ;
+        sh:description """Check if the RO-Crate Metadata File Descriptor entity exists,
+            i.e., if there exists an entity with @id `ro-crate-metadata.json` and type `schema:CreativeWork`""" ;
+        sh:message """The root of the document MUST have an entity with @id `ro-crate-metadata.json`, typed as `schema:CreativeWork` and with an `about` property referencing the Root Data Entity""" ;
+        sh:prefixes ro-crate:sparqlPrefixes ;
+        sh:select """
+            SELECT ?this
+            WHERE {
+                FILTER NOT EXISTS {
+                    ?crate a schema:Dataset .
+                    ?md schema:about ?crate .
+                    ?md a schema:CreativeWork .
+                    FILTER( STRENDS(STR(?md), "ro-crate-metadata.json") )
+                }
+            }
+        """ ;
+        ] .
 
-ro-crate:FindROCrateMetadataFileDescriptorEntity a sh:NodeShape, validator:HiddenShape;
+ro-crate:FindROCrateMetadataFileDescriptorEntity a sh:NodeShape, validator:HiddenShape ;
     sh:name "Identify the RO-Crate Metadata File Descriptor" ;
     sh:description """The RO-Crate Metadata File Descriptor entity describes the RO-Crate itself, and it is named as `ro-crate-metadata.json`.
         It can be identified by name according to the RO-Crate specification
@@ -32,7 +54,8 @@ ro-crate:FindROCrateMetadataFileDescriptorEntity a sh:NodeShape, validator:Hidde
         sh:select """
             SELECT ?this
             WHERE {
-                ?this a schema:CreativeWork ;
+                ?crate a schema:Dataset .
+                ?this schema:about ?crate .
                 FILTER(contains(str(?this), "ro-crate-metadata.json"))
             }
         """
@@ -46,29 +69,12 @@ ro-crate:FindROCrateMetadataFileDescriptorEntity a sh:NodeShape, validator:Hidde
         sh:object ro-crate:ROCrateMetadataFileDescriptor ;
     ] .
 
-ro-crate:ROCrateMetadataFileDescriptorExistence
-    a sh:NodeShape ;
-    sh:name "RO-Crate Metadata File Descriptor entity existence" ;
-    sh:description "The RO-Crate JSON-LD MUST contain a Metadata File Descriptor entity named `ro-crate-metadata.json` and typed as `schema:CreativeWork`" ;
-    sh:targetNode ro:ro-crate-metadata.json ;
-    sh:property [
-        a sh:PropertyShape ;
-        sh:name "RO-Crate Metadata File Descriptor entity existence" ;
-        sh:description """Check if the RO-Crate Metadata File Descriptor entity exists,
-            i.e., if there exists an entity with @id `ro-crate-metadata.json` and type `schema:CreativeWork`""" ;
-        sh:path rdf:type ;
-        sh:hasValue ro-crate:ROCrateMetadataFileDescriptor ;
-        sh:minCount 1 ;
-        sh:message "The root of the document MUST have an entity with @id `ro-crate-metadata.json`" ;
-    ] .
-
 ro-crate:ROCrateMetadataFileDescriptorRecommendedProperties a sh:NodeShape ;
     sh:name "RO-Crate Metadata File Descriptor REQUIRED properties" ;
     sh:description """RO-Crate Metadata Descriptor MUST be defined
-    according with the requirements details defined in
-    [RO-Crate Metadata File Descriptor](https://www.researchobject.org/ro-crate/1.1/root-data-entity.html#ro-crate-metadata-file-descriptor)""";
-    sh:targetNode ro:ro-crate-metadata.json ;
-    sh:property [ 
+    according with the requirements details defined in [RO-Crate Metadata File Descriptor](https://www.researchobject.org/ro-crate/1.1/root-data-entity.html#ro-crate-metadata-file-descriptor)""" ;
+    sh:targetClass ro-crate:ROCrateMetadataFileDescriptor ;
+    sh:property [
         a sh:PropertyShape ;
         sh:name "Metadata File Descriptor entity type" ;
         sh:description "Check if the RO-Crate Metadata File Descriptor has `@type` CreativeWork, as per schema.org" ;
@@ -78,11 +84,11 @@ ro-crate:ROCrateMetadataFileDescriptorRecommendedProperties a sh:NodeShape ;
         sh:hasValue schema_org:CreativeWork ;
         sh:message "The RO-Crate metadata file MUST be a CreativeWork, as per schema.org" ;
     ] ;
-    sh:property [ 
+    sh:property [
         a sh:PropertyShape ;
         sh:name "Metadata File Descriptor entity: `about` property" ;
         sh:description """Check if the RO-Crate Metadata File Descriptor has an `about` property referencing the Root Data Entity""" ;
-        sh:maxCount 1;
+        sh:maxCount 1 ;
         sh:minCount 1 ;
         sh:nodeKind sh:IRI ;
         sh:path schema_org:about ;

rocrate_validator/profiles/ro-crate/ontology.ttl

@@ -31,6 +31,10 @@
 # # #    Classes
 # # #################################################################
 
+# Declare the ROCrateMetadataFileDescriptor class
+ro-crate:ROCrateMetadataFileDescriptor rdf:type owl:Class ;
+  rdfs:label "ROCrateMetadataFileDescriptor"@en .
+
 # Declare the RootDataEntity class
 ro-crate:RootDataEntity rdf:type owl:Class ;
   rdfs:subClassOf schema:Dataset ;
@@ -65,3 +69,12 @@ ro-crate:DataEntity rdf:type owl:Class ;
 ro-crate:Directory rdf:type owl:Class ;
                 rdfs:subClassOf schema:Dataset ;
                 rdfs:label "Directory"@en .
+
+
+# # #################################################################
+# # #    Individuals
+# # ##############################################################
+
+# Declare the individual representing the RO-Crate Metadata File Descriptor entity
+ro:ro-crate-metadata.json a owl:NamedIndividual, ro-crate:ROCrateMetadataFileDescriptor .
+

rocrate_validator/requirements/shacl/validator.py

@@ -23,7 +23,6 @@
 from rdflib import BNode, Graph
 from rdflib.term import Node, URIRef
 
-from rocrate_validator.utils import log as logging
 from rocrate_validator.constants import (DEFAULT_ONTOLOGY_FILE,
                                          RDF_SERIALIZATION_FORMATS,
                                          RDF_SERIALIZATION_FORMATS_TYPES,
@@ -34,6 +33,8 @@
 from rocrate_validator.requirements.shacl.models import ShapesRegistry
 from rocrate_validator.requirements.shacl.utils import (make_uris_relative,
                                                         map_severity)
+from rocrate_validator.utils import log as logging
+from rocrate_validator.utils.rdf import extract_base_from_jsonld
 
 # set up logging
 logger = logging.getLogger(__name__)
@@ -189,6 +190,22 @@ def __get_ontology_path__(self, profile_path: Path, ontology_filename: str = DEF
             self._ontology_path = Path(f"{profile_path}/{ontology_filename}")
         return self._ontology_path
 
+    def __get_data_graph_base__(self) -> Optional[str]:
+        """
+        Get the @base from the RO-Crate metadata JSON-LD.
+
+        This extracts the @base from the @context of the data graph metadata,
+        which can be used to align the ontology graph's base URI with the data graph.
+
+        :return: The @base value if found, None otherwise
+        """
+        try:
+            metadata_dict = self.ro_crate.metadata.as_dict()
+            return extract_base_from_jsonld(metadata_dict)
+        except Exception as e:
+            logger.debug("Unable to extract @base from data graph metadata: %s", e)
+            return None
+
     def __load_ontology_graph__(self, profile_path: Path,
                                 ontology_filename: Optional[str] = DEFAULT_ONTOLOGY_FILE) -> Graph:
         # load the graph of ontologies
@@ -197,8 +214,20 @@ def __load_ontology_graph__(self, profile_path: Path,
         if os.path.exists(ontology_path):
             logger.debug("Loading ontologies: %s", ontology_path)
             ontology_graph = Graph()
+
+            # Determine the publicID to use:
+            # 1. First, try to get @base from the data graph metadata
+            # 2. Fall back to the default publicID (RO-Crate URI)
+            data_graph_base = self.__get_data_graph_base__()
+            public_id = data_graph_base if data_graph_base else self.publicID
+
+            if data_graph_base:
+                logger.debug("Using @base from data graph metadata: %s", data_graph_base)
+            else:
+                logger.debug("Using default publicID: %s", self.publicID)
+
             ontology_graph.parse(ontology_path, format="ttl",
-                                 publicID=self.publicID)
+                                 publicID=public_id)
             logger.debug("Ontologies loaded: %s", ontology_graph)
         return ontology_graph
 

rocrate_validator/utils/rdf.py

@@ -11,12 +11,16 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+from typing import Optional
 
 from rdflib import Graph
-from rocrate_validator.utils import logger
-from rocrate_validator.utils.paths import list_graph_paths
 
 from rocrate_validator import constants
+from rocrate_validator.utils import log as logging
+from rocrate_validator.utils.paths import list_graph_paths
+
+# set up logging
+logger = logging.getLogger(__name__)
 
 
 def get_full_graph(
@@ -37,3 +41,32 @@ def get_full_graph(
         full_graph.parse(graph_path, format="turtle", publicID=publicID)
         logger.debug("Loaded triples from %s", graph_path)
     return full_graph
+
+
+def extract_base_from_jsonld(json_data: dict) -> Optional[str]:
+    """
+    Extract the @base from the @context of a JSON-LD document.
+
+    The @context can be:
+    - A dictionary (e.g., {"@base": "http://example.org/"})
+    - A list of contexts (e.g., [{"@base": "http://example.org/"}, "https://schema.org"])
+
+    :param json_data: The JSON-LD data as a dictionary
+    :return: The @base value if found, None otherwise
+    """
+    context = json_data.get('@context')
+
+    if not context:
+        return None
+
+    # If @context is a dictionary, look for @base directly
+    if isinstance(context, dict):
+        return context.get('@base')
+
+    # If @context is a list, look for @base in each context item
+    if isinstance(context, list):
+        for ctx in context:
+            if isinstance(ctx, dict) and '@base' in ctx:
+                return ctx['@base']
+
+    return None