Adding option for non-redundant entialed relationships. (#740)

* Adding option for non-redundant entialed relationships. Fixes #739 * tests * removing ---s with messed with CLI formattinh * Extended relationships notebook for entailed-direct * skip correct test
INCATools · Apr 25, 2024 · 64c6114 · 64c6114
1 parent a52eed5
commit 64c6114
Show file tree

Hide file tree

Showing 9 changed files with 2,611 additions and 670 deletions.
diff --git a/notebooks/Commands/Relationships.ipynb b/notebooks/Commands/Relationships.ipynb
diff --git a/notebooks/Commands/ValidateDefinitions.ipynb b/notebooks/Commands/ValidateDefinitions.ipynb
diff --git a/notebooks/Commands/output/trigeminal-ganglion-graph.png b/notebooks/Commands/output/trigeminal-ganglion-graph.png
diff --git a/src/oaklib/cli.py b/src/oaklib/cli.py
diff --git a/src/oaklib/interfaces/obograph_interface.py b/src/oaklib/interfaces/obograph_interface.py
@@ -259,6 +259,38 @@ def descendant_graph(
             self.transitive_query_cache[key] = g
         return g
 
+    def non_redundant_entailed_relationships(
+        self,
+        predicates: List[PRED_CURIE] = None,
+        **kwargs,
+    ) -> Iterator[RELATIONSHIP]:
+        """
+        Yields all relationships that are directly entailed.
+
+        See https://github.com/INCATools/ontology-access-kit/issues/739
+
+        :param kwargs: same as relationships
+        :return:
+        """
+        if "include_entailed" in kwargs:
+            kwargs.pop("include_entailed")
+        relationships = list(
+            self.relationships(predicates=predicates, include_entailed=True, **kwargs)
+        )
+        rel_by_sp = defaultdict(list)
+        for s, p, o in relationships:
+            if s == o:
+                continue
+            rel_by_sp[(s, p)].append(o)
+        for (s, p), objs in rel_by_sp.items():
+            redundant_set = set()
+            for o in objs:
+                ancs = list(self.ancestors(o, predicates=predicates, reflexive=False))
+                redundant_set.update(ancs)
+            for o in objs:
+                if o not in redundant_set:
+                    yield s, p, o
+
     def ancestors(
         self,
         start_curies: Union[CURIE, List[CURIE]],

diff --git a/tests/__init__.py b/tests/__init__.py
@@ -46,6 +46,8 @@ def output_path(fn: str) -> str:
 NUCLEUS = "GO:0005634"
 ORGANELLE_MEMBRANE = "GO:0031090"
 NUCLEAR_ENVELOPE = "GO:0005635"
+ORGANELLE_ENVELOPE = "GO:0031967"
+ENVELOPE = "GO:0031975"
 THYLAKOID = "GO:0009579"
 ATOM = "CHEBI:33250"
 INTERNEURON = "CL:0000099"

diff --git a/tests/test_implementations/__init__.py b/tests/test_implementations/__init__.py
@@ -96,6 +96,7 @@
     CYTOPLASM,
     CYTOPLASMIC_REGION,
     ENDOMEMBRANE_SYSTEM,
+    ENVELOPE,
     EUKARYOTA,
     FAKE_ID,
     FUNGI,
@@ -120,6 +121,8 @@
     NUCLEUS,
     OPISTHOKONTA,
     ORGANELLE,
+    ORGANELLE_ENVELOPE,
+    ORGANELLE_MEMBRANE,
     PHENOTYPIC_ABNORMALITY,
     PHOTORECEPTOR_OUTER_SEGMENT,
     PHOTOSYNTHETIC_MEMBRANE,
@@ -615,6 +618,42 @@ def test_relationships(self, oi: BasicOntologyInterface, ignore_annotation_edges
                 irels = list(oi.incoming_relationships(o, predicates=[p]))
                 test.assertIn((p, s), irels)
 
+    def test_entailed_relationships(self, oi: OboGraphInterface):
+        """
+        Tests entailed relationship methods for compliance.
+
+        :param oi:
+        :return:
+        """
+        test = self.test
+        cases = [
+            (
+                NUCLEAR_MEMBRANE,
+                [IS_A, PART_OF],
+                {IS_A: {ORGANELLE_MEMBRANE}, PART_OF: {NUCLEAR_ENVELOPE}},
+            ),
+            (
+                NUCLEAR_MEMBRANE,
+                [IS_A, OVERLAPS],
+                {IS_A: {ORGANELLE_MEMBRANE}, OVERLAPS: {NUCLEAR_ENVELOPE}},
+            ),
+            (NUCLEAR_MEMBRANE, [IS_A], {IS_A: {ORGANELLE_MEMBRANE}}),
+            (
+                NUCLEAR_MEMBRANE,
+                [PART_OF],
+                {PART_OF: {NUCLEAR_ENVELOPE, ORGANELLE_ENVELOPE, ENVELOPE}},
+            ),
+        ]
+        for curie, preds, expected in cases:
+            logging.info(f"TESTS FOR {curie}")
+            rels = list(oi.non_redundant_entailed_relationships(subjects=[curie], predicates=preds))
+            objs_by_pred = {p: set() for p in preds}
+            for s, p, o in rels:
+                objs_by_pred[p].add(o)
+                assert s == curie
+            for p in preds:
+                test.assertCountEqual(expected[p], objs_by_pred[p])
+
     def test_rbox_relationships(self, oi: BasicOntologyInterface):
         """
         Tests relationships between relationship types

diff --git a/tests/test_implementations/test_simple_obo.py b/tests/test_implementations/test_simple_obo.py
@@ -114,6 +114,10 @@ def test_relationships_extra(self):
     def test_relationships(self):
         self.compliance_tester.test_relationships(self.oi)
 
+    @unittest.skip("Contents of go-nucleus file need to be aligned")
+    def test_entailed_relationships(self):
+        self.compliance_tester.test_entailed_relationships(self.oi)
+
     def test_rbox_relationships(self):
         self.compliance_tester.test_rbox_relationships(self.oi)
 

diff --git a/tests/test_implementations/test_sqldb.py b/tests/test_implementations/test_sqldb.py
@@ -90,6 +90,10 @@ def test_relationships(self):
         oi = SqlImplementation(OntologyResource(slug=f"sqlite:///{str(DB)}"))
         self.compliance_tester.test_relationships(oi, ignore_annotation_edges=False)
 
+    def test_entailed_relationships(self):
+        oi = SqlImplementation(OntologyResource(slug=f"sqlite:///{str(DB)}"))
+        self.compliance_tester.test_entailed_relationships(oi)
+
     def test_relationships_chunking(self):
         """
         Tests behavior for chunking relationship queries