fix(shacl): extract @base from JSON-LD for ontology parsing

Author: kikkomep

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