Supporting skos mappings for simpleobo and pronto. (#766)

* Supporting skos mappings for simpleobo and pronto.

* format

* add missing

* add missing

src/oaklib/implementations/pronto/pronto_implementation.py

@@ -40,7 +40,6 @@
     OWL_VERSION_INFO,
     SCOPE_TO_SYNONYM_PRED_MAP,
     SEMAPV,
-    SKOS_CLOSE_MATCH,
     SKOS_MATCH_PREDICATES,
     TERM_REPLACED_BY,
     TERMS_MERGED,
@@ -549,6 +548,12 @@ def simple_mappings_by_curie(self, curie: CURIE) -> Iterable[Tuple[PRED_CURIE, C
                     yield pred, v
                 elif isinstance(s, ResourcePropertyValue):
                     yield pred, s.resource
+        if isinstance(t, Term):
+            for rel_type, parents in t.relationships.items():
+                pred = self._get_pronto_relationship_type_curie(rel_type)
+                if pred in SKOS_MATCH_PREDICATES:
+                    for parent in parents:
+                        yield pred, parent.id
 
     def entity_metadata_map(self, curie: CURIE) -> METADATA_MAP:
         t = self._entity(curie)
@@ -662,7 +667,7 @@ def sssom_mappings(
                     if x.id in curies:
                         m = sssom.Mapping(
                             subject_id=e,
-                            predicate_id=SKOS_CLOSE_MATCH,
+                            predicate_id=HAS_DBXREF,
                             object_id=x.id,
                             mapping_justification=sssom.EntityReference(
                                 SEMAPV.UnspecifiedMatching.value

src/oaklib/implementations/simpleobo/simple_obo_implementation.py

@@ -59,7 +59,7 @@
     RDFS_RANGE,
     SCOPE_TO_SYNONYM_PRED_MAP,
     SEMAPV,
-    SKOS_CLOSE_MATCH,
+    SKOS_MATCH_PREDICATES,
     SUBPROPERTY_OF,
     TERM_REPLACED_BY,
     TERMS_MERGED,
@@ -697,20 +697,46 @@ def get_sssom_mappings_by_curie(self, curie: Union[str, CURIE]) -> Iterator[ssso
             for x in s.simple_values(TAG_XREF):
                 m = sssom.Mapping(
                     subject_id=curie,
-                    predicate_id=SKOS_CLOSE_MATCH,
+                    predicate_id=HAS_DBXREF,
                     object_id=x,
                     mapping_justification=sssom.EntityReference(SEMAPV.UnspecifiedMatching.value),
                 )
                 inject_mapping_sources(m)
                 yield m
+            for x in s.property_values():
+                p = self.map_shorthand_to_curie(x[0])
+                if p in SKOS_MATCH_PREDICATES:
+                    m = sssom.Mapping(
+                        subject_id=curie,
+                        predicate_id=p,
+                        object_id=x[1],
+                        mapping_justification=sssom.EntityReference(
+                            SEMAPV.UnspecifiedMatching.value
+                        ),
+                    )
+                    inject_mapping_sources(m)
+                    yield m
+            for p, v in s.pair_values(TAG_RELATIONSHIP):
+                p = self.map_shorthand_to_curie(p)
+                if p in SKOS_MATCH_PREDICATES:
+                    m = sssom.Mapping(
+                        subject_id=curie,
+                        predicate_id=p,
+                        object_id=v,
+                        mapping_justification=sssom.EntityReference(
+                            SEMAPV.UnspecifiedMatching.value
+                        ),
+                    )
+                    inject_mapping_sources(m)
+                    yield m
         # TODO: use a cache to avoid re-calculating
         for _, stanza in self.obo_document.stanzas.items():
             if len(stanza.simple_values(TAG_XREF)) > 0:
                 for x in stanza.simple_values(TAG_XREF):
                     if x == curie:
                         m = sssom.Mapping(
                             subject_id=stanza.id,
-                            predicate_id=SKOS_CLOSE_MATCH,
+                            predicate_id=HAS_DBXREF,
                             object_id=curie,
                             mapping_justification=SEMAPV.UnspecifiedMatching.value,
                         )

tests/input/mapping-predicates-test.obo

@@ -0,0 +1,30 @@
+format-version: 1.2
+ontology: test.obo
+idspace: dc http://purl.org/dc/elements/1.1/ 
+idspace: oboInOwl http://www.geneontology.org/formats/oboInOwl# 
+idspace: owl http://www.w3.org/2002/07/owl# 
+idspace: rdf http://www.w3.org/1999/02/22-rdf-syntax-ns# 
+idspace: rdfs http://www.w3.org/2000/01/rdf-schema# 
+idspace: terms http://purl.org/dc/terms/ 
+idspace: xml http://www.w3.org/XML/1998/namespace 
+idspace: xsd http://www.w3.org/2001/XMLSchema# 
+idspace: X http://purl.obolibrary.org/obo/X_
+
+[Term]
+id: X:0000001
+name: Person
+property_value: skos:exactMatch schema:Person
+relationship: skos:closeMatch prov:Agent
+xref: Y:1
+
+[Typedef]
+id: skos:exactMatch
+is_metadata_tag: true
+is_class_level: true
+
+
+[Typedef]
+id: skos:closeMatch
+is_metadata_tag: true
+is_class_level: true
+

tests/test_implementations/__init__.py

@@ -458,6 +458,19 @@ def test_multilingual(self, oi: BasicOntologyInterface):
                     f"Definition for {lang} did not match",
                 )
 
+    def test_skos_mappings(self, oi: MappingProviderInterface):
+        for curies in [None, ["X:0000001"]]:
+            mappings = [
+                (m.subject_id, m.predicate_id, m.object_id) for m in oi.sssom_mappings(curies)
+            ]
+            print(mappings)
+            expected = [
+                ("X:0000001", "oio:hasDbXref", "Y:1"),
+                ("X:0000001", "skos:exactMatch", "schema:Person"),
+                ("X:0000001", "skos:closeMatch", "prov:Agent"),
+            ]
+            self.test.assertCountEqual(expected, mappings)
+
     def test_sssom_mappings(self, oi: MappingProviderInterface):
         """
         Tests conformance of MappingProviderInterface.

tests/test_implementations/test_pronto.py

@@ -44,6 +44,7 @@
 TEST_SIMPLE_ONT = INPUT_DIR / "go-nucleus-simple.obo"
 TEST_ONT_COPY = OUTPUT_DIR / "go-nucleus.copy.obo"
 TEST_SUBGRAPH_OUT = OUTPUT_DIR / "vacuole.obo"
+TEST_SKOS_MAPPINGS_ONT = INPUT_DIR / "mapping-predicates-test.obo"
 
 
 class TestProntoImplementation(unittest.TestCase):
@@ -193,6 +194,16 @@ def test_obsolete_entities(self):
         oi = ProntoImplementation(resource)
         self.compliance_tester.test_obsolete_entities(oi)
 
+    @unittest.skip("Pronto does not handling dangling references")
+    def test_skos_mappings(self):
+        """
+        Tests mappings as SKOS properties.
+
+        :return:
+        """
+        adapter = get_adapter(f"pronto:{TEST_SKOS_MAPPINGS_ONT}")
+        self.compliance_tester.test_skos_mappings(adapter)
+
     def test_sssom_mappings(self):
         self.compliance_tester.test_sssom_mappings(self.oi)
 

tests/test_implementations/test_simple_obo.py

@@ -3,6 +3,7 @@
 from copy import deepcopy
 
 from kgcl_schema.datamodel import kgcl
+from oaklib import get_adapter
 from oaklib.datamodels import obograph
 from oaklib.datamodels.search import SearchConfiguration
 from oaklib.datamodels.search_datamodel import SearchProperty, SearchTermSyntax
@@ -59,6 +60,7 @@
 TEST_SIMPLE_ONT = INPUT_DIR / "go-nucleus-simple.obo"
 TEST_ONT_COPY = OUTPUT_DIR / "go-nucleus.copy.obo"
 TEST_SUBGRAPH_OUT = OUTPUT_DIR / "vacuole.obo"
+TEST_SKOS_MAPPINGS_ONT = INPUT_DIR / "mapping-predicates-test.obo"
 
 
 class TestSimpleOboImplementation(unittest.TestCase):
@@ -236,6 +238,15 @@ def test_synonyms_extra(self):
     def test_sssom_mappings(self):
         self.compliance_tester.test_sssom_mappings(self.oi)
 
+    def test_skos_mappings(self):
+        """
+        Tests mappings as SKOS properties.
+
+        :return:
+        """
+        adapter = get_adapter(f"simpleobo:{TEST_SKOS_MAPPINGS_ONT}")
+        self.compliance_tester.test_skos_mappings(adapter)
+
     def test_definitions(self):
         self.compliance_tester.test_definitions(self.oi, include_metadata=True)