Fix bibliographic

app/models/profiles/profile.py

@@ -1,5 +1,6 @@
 from abc import ABC, abstractmethod
 from pathlib import Path
+import json
 
 from jinja2 import Environment, FileSystemLoader, select_autoescape
 from lxml import etree
@@ -53,6 +54,11 @@ def shacl_sip() -> Path:
     def shacl_profile() -> Path:
         pass
 
+    @staticmethod
+    @abstractmethod
+    def shacl_ie() -> Path:
+        pass
+
     @staticmethod
     @abstractmethod
     def profile_name() -> str:
@@ -177,9 +183,21 @@ def _construct_shacl_graph(self) -> Graph:
         Returns:
             A SHACL graph.
         """
-
         pass
 
+    def _construct_ie_shacl_graph(self) -> Graph:
+        """Construct a graph containing the count Intellectual Entity SHACL.
+
+        This is used for validating the data graph.
+
+        Returns:
+            A SHACL graph.
+        """
+
+        shacl_graph = Graph()
+        shacl_graph.parse(str(self.shacl_ie()), format="turtle")
+        return shacl_graph
+
     def validate_graph(self, data_graph: Graph) -> bool:
         """Validate if the graph is conform.
 
@@ -198,6 +216,37 @@ def validate_graph(self, data_graph: Graph) -> bool:
             raise GraphNotConformError(
                 "Graph is perceived as empty as it does not contain an intellectual entity."
             )
+        result_query = """
+                PREFIX sh: <http://www.w3.org/ns/shacl#> 
+                PREFIX haObjId: <https://data.hetarchief.be/id/entity/>
+                SELECT ?message ?severity ?focusNode
+                WHERE {
+                    ?s sh:resultMessage ?message ;
+                    sh:resultSeverity ?severity_iri ;
+                    sh:focusNode ?focusNode_iri
+                    BIND(REPLACE(str(?severity_iri), str(sh:), "") as ?severity)
+                    BIND(REPLACE(str(?focusNode_iri), str(haObjId:), "") as ?focusNode)
+                } ORDER BY ?severity
+        """
+        ie_shacl_graph = self._construct_ie_shacl_graph()
+        conforms, results_graph, results_text = shacl_validate(
+            data_graph=data_graph,
+            shacl_graph=ie_shacl_graph,
+            meta_shacl=True,
+            allow_warnings=True,
+        )
+
+        if not conforms:
+            results = json.loads(
+                results_graph.query(result_query).serialize(format="json")
+            )
+            result_formatted_text = ""
+            for result in results["results"]["bindings"]:
+                result_formatted_text += f"Severity: {result['severity']['value']}\n"
+                result_formatted_text += f"Id: {result['focusNode']['value']}\n"
+                result_formatted_text += f"Message: {result['message']['value']}\n\n"
+            raise GraphNotConformError(result_formatted_text)
+
         shacl_graph = self._construct_shacl_graph()
         conforms, results_graph, results_text = shacl_validate(
             data_graph=data_graph,
@@ -207,6 +256,14 @@ def validate_graph(self, data_graph: Graph) -> bool:
         )
 
         if not conforms:
-            raise GraphNotConformError(results_text)
+            results = json.loads(
+                results_graph.query(result_query).serialize(format="json")
+            )
+            result_formatted_text = ""
+            for result in results["results"]["bindings"]:
+                result_formatted_text += f"Severity: {result['severity']['value']}\n"
+                result_formatted_text += f"Id: {result['focusNode']['value']}\n"
+                result_formatted_text += f"Message: {result['message']['value']}\n\n"
+            raise GraphNotConformError(result_formatted_text)
 
         return True

app/models/profiles/profile_1_1.py

@@ -110,6 +110,17 @@ def dcterms_xsd() -> Path:
             "dc_basic.xsd",
         )
 
+    @staticmethod
+    def shacl_ie() -> Path:
+        return Path(
+            "app",
+            "resources",
+            "1.1",
+            "basic",
+            "shacl",
+            "intellectual_entity_count.shacl.ttl",
+        )
+
     def _validate_descriptive(self) -> list[XMLNotValidError]:
         """Validate the dcterms file.
 
@@ -340,6 +351,17 @@ def dc_schema_xsd() -> Path:
             "descriptive_material_artwork.xsd",
         )
 
+    @staticmethod
+    def shacl_ie() -> Path:
+        return Path(
+            "app",
+            "resources",
+            "1.1",
+            "material_artwork",
+            "shacl",
+            "intellectual_entity_count.shacl.ttl",
+        )
+
     def _validate_descriptive(self) -> list[XMLNotValidError]:
         """Validate the dc+schema files.
 
@@ -669,6 +691,17 @@ def shacl_premis() -> Path:
             "premis.shacl.ttl",
         )
 
+    @staticmethod
+    def shacl_ie() -> Path:
+        return Path(
+            "app",
+            "resources",
+            "1.1",
+            "newspaper",
+            "shacl",
+            "intellectual_entity_count.shacl.ttl",
+        )
+
     def _validate_descriptive(self) -> list[XMLNotValidError]:
         """Validate the MODS or dc+schema files.
 

app/models/profiles/profile_1_2.py

@@ -2,8 +2,7 @@
 from lxml import etree
 from rdflib import Graph
 
-from app.models.profiles.profile import sa
-from app.models.profiles.exceptions import XMLNotValidError, GraphParseError
+from app.models.profiles.exceptions import XMLNotValidError
 from app.models.profiles.profile_1_1 import (
     BasicProfile11,
     NewspaperProfile11,
@@ -129,6 +128,17 @@ def dc_schema_xsd() -> Path:
             "descriptive_basic.xsd",
         )
 
+    @staticmethod
+    def shacl_ie() -> Path:
+        return Path(
+            "app",
+            "resources",
+            "1.2",
+            "basic",
+            "shacl",
+            "intellectual_entity_count.shacl.ttl",
+        )
+
     def _validate_descriptive(self) -> list[XMLNotValidError]:
         """Validate the dc+schema file.
 
@@ -288,6 +298,17 @@ def dc_schema_xsd() -> Path:
             "descriptive_material_artwork.xsd",
         )
 
+    @staticmethod
+    def shacl_ie() -> Path:
+        return Path(
+            "app",
+            "resources",
+            "1.2",
+            "material_artwork",
+            "shacl",
+            "intellectual_entity_count.shacl.ttl",
+        )
+
 
 class BibliographicProfile12(NewspaperProfile11):
     def profile_name() -> str:
@@ -424,3 +445,14 @@ def shacl_premis() -> Path:
             "shacl",
             "premis.shacl.ttl",
         )
+
+    @staticmethod
+    def shacl_ie() -> Path:
+        return Path(
+            "app",
+            "resources",
+            "1.2",
+            "bibliographic",
+            "shacl",
+            "intellectual_entity_count.shacl.ttl",
+        )

app/resources/1.1/basic/shacl/intellectual_entity_count.shacl.ttl

@@ -0,0 +1,21 @@
+@prefix premis: <http://www.loc.gov/premis/rdf/v3/> .
+@prefix sh: <http://www.w3.org/ns/shacl#> .
+@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
+
+<#IntellectualEntityCountShape>
+  a sh:NodeShape ;
+  sh:targetNode premis:IntellectualEntity ;
+  sh:property [
+    sh:path [ sh:inversePath rdf:type ] ;
+    sh:minCount 1 ;
+    sh:severity sh:Violation ;
+    sh:message """No Intellectual Entity found.
+Please check `premis:premis/premis:object/premis:objectIdentifier` and `mods:mods/mods:identifier[not(@type)]`"""@en ;
+  ],
+  [
+    sh:path [ sh:inversePath rdf:type ] ;
+    sh:maxCount 1 ;
+    sh:severity sh:Violation ;
+    sh:message """More than one Intellectual Entity found.
+Please check if `premis:premis/premis:object/premis:objectIdentifier` and `mods:mods/mods:identifier[not(@type)]` have the same value."""@en ;
+  ] .

app/resources/1.1/material_artwork/shacl/intellectual_entity_count.shacl.ttl

@@ -0,0 +1,21 @@
+@prefix premis: <http://www.loc.gov/premis/rdf/v3/> .
+@prefix sh: <http://www.w3.org/ns/shacl#> .
+@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
+
+<#IntellectualEntityCountShape>
+  a sh:NodeShape ;
+  sh:targetNode premis:IntellectualEntity ;
+  sh:property [
+    sh:path [ sh:inversePath rdf:type ] ;
+    sh:minCount 1 ;
+    sh:severity sh:Violation ;
+    sh:message """No Intellectual Entity found.
+Please check `premis:premis/premis:object/premis:objectIdentifier` and `mods:mods/mods:identifier[not(@type)]`"""@en ;
+  ],
+  [
+    sh:path [ sh:inversePath rdf:type ] ;
+    sh:maxCount 1 ;
+    sh:severity sh:Violation ;
+    sh:message """More than one Intellectual Entity found.
+Please check if `premis:premis/premis:object/premis:objectIdentifier` and `mods:mods/mods:identifier[not(@type)]` have the same value."""@en ;
+  ] .

app/resources/1.1/newspaper/shacl/intellectual_entity_count.shacl.ttl

@@ -0,0 +1,21 @@
+@prefix premis: <http://www.loc.gov/premis/rdf/v3/> .
+@prefix sh: <http://www.w3.org/ns/shacl#> .
+@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
+
+<#IntellectualEntityCountShape>
+  a sh:NodeShape ;
+  sh:targetNode premis:IntellectualEntity ;
+  sh:property [
+    sh:path [ sh:inversePath rdf:type ] ;
+    sh:minCount 1 ;
+    sh:severity sh:Violation ;
+    sh:message """No Intellectual Entity found.
+Please check `premis:premis/premis:object/premis:objectIdentifier` and `mods:mods/mods:identifier[not(@type)]`"""@en ;
+  ],
+  [
+    sh:path [ sh:inversePath rdf:type ] ;
+    sh:maxCount 1 ;
+    sh:severity sh:Violation ;
+    sh:message """More than one Intellectual Entity found.
+Please check if `premis:premis/premis:object/premis:objectIdentifier` and `mods:mods/mods:identifier[not(@type)]` have the same value."""@en ;
+  ] .

app/resources/1.1/newspaper/shacl/premis.shacl.ttl

@@ -187,9 +187,20 @@
     sh:nodeKind sh:IRI ;
     sh:class premis:File ;
     sh:name "is source of"@en ;
-    sh:description "The source of the file"@en ;
+    sh:description "Is the source of the file"@en ;
     sh:message """The File's value for `is source of` relationshipSubType is not a File.
-      Please check `premis:premis/premis:object[@xsi:type=premis:File]/premis:relationship/(premis:relationshipSubType="is included in")`
+      Please check `premis:premis/premis:object[@xsi:type=premis:File]/premis:relationship/(premis:relationshipSubType="is source of")`
+      OR check for mismatch between the File ids `premis:premis/premis:object/premis:objectIdentifier/premis:objectIdentifierValue`."""@en ;
+  ],
+  [
+    a sh:PropertyShape ;
+    sh:path relSubType:hss ;
+    sh:nodeKind sh:IRI ;
+    sh:class premis:File ;
+    sh:name "has source"@en ;
+    sh:description "The source of the file"@en ;
+    sh:message """The File's value for `has source` relationshipSubType is not a File.
+      Please check `premis:premis/premis:object[@xsi:type=premis:File]/premis:relationship/(premis:relationshipSubType="has source")`
       OR check for mismatch between the File ids `premis:premis/premis:object/premis:objectIdentifier/premis:objectIdentifierValue`."""@en ;
   ],
   [

app/resources/1.2/basic/shacl/intellectual_entity_count.shacl.ttl

@@ -0,0 +1,21 @@
+@prefix premis: <http://www.loc.gov/premis/rdf/v3/> .
+@prefix sh: <http://www.w3.org/ns/shacl#> .
+@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
+
+<#IntellectualEntityCountShape>
+  a sh:NodeShape ;
+  sh:targetNode premis:IntellectualEntity ;
+  sh:property [
+    sh:path [ sh:inversePath rdf:type ] ;
+    sh:minCount 1 ;
+    sh:severity sh:Violation ;
+    sh:message """No Intellectual Entity found.
+Please check `premis:premis/premis:object/premis:objectIdentifier` and `mods:mods/mods:identifier[not(@type)]`"""@en ;
+  ],
+  [
+    sh:path [ sh:inversePath rdf:type ] ;
+    sh:maxCount 1 ;
+    sh:severity sh:Violation ;
+    sh:message """More than one Intellectual Entity found.
+Please check if `premis:premis/premis:object/premis:objectIdentifier` and `mods:mods/mods:identifier[not(@type)]` have the same value."""@en ;
+  ] .