Fixing missing OWL label as alias bug, see #52

pyproject.toml

@@ -25,7 +25,7 @@ jupyter = "^1.0.0"
 sphinx-rtd-theme = "^1.0.0"
 sphinx-click = "^3.1.0"
 myst-parser = "^0.17.0"
-linkml = "^1.2.4"
+linkml = "^1.2.10"
 sphinxcontrib-mermaid = "^0.7.1"
 coverage = "^6.3.2"
 

src/oaklib/implementations/sparql/abstract_sparql_implementation.py

@@ -284,7 +284,7 @@ def _alias_predicates(self) -> List[PRED_CURIE]:
 
     def alias_map_by_curie(self, curie: CURIE) -> ALIAS_MAP:
         uri = self.curie_to_sparql(curie)
-        alias_pred_uris = [self.curie_to_sparql(p) for p in self._alias_predicates()]
+        alias_pred_uris = [self.curie_to_sparql(p) for p in self._alias_predicates() + [LABEL_PREDICATE]]
         query = SparqlQuery(select=['?p', '?o'],
                             where=[f'{uri} ?p ?o',
                                    _sparql_values('p', alias_pred_uris)])
@@ -294,7 +294,7 @@ def alias_map_by_curie(self, curie: CURIE) -> ALIAS_MAP:
                                {'oboInOwl': 'http://www.geneontology.org/formats/oboInOwl#'})
         m = defaultdict(list)
         for row in bindings:
-            m[row['p']['value']].append(row['o']['value'])
+            m[self.uri_to_curie(row['p']['value'])].append(row['o']['value'])
         return m
 
     def get_curies_by_label(self, label: str) -> List[CURIE]:

src/oaklib/interfaces/basic_ontology_interface.py

@@ -289,6 +289,9 @@ def alias_map_by_curie(self, curie: CURIE) -> ALIAS_MAP:
         """
         Returns aliases keyed by alias type (scope in OBO terms)
 
+        - The alias map MUST include rdfs:label annotations
+        - The alias map MAY include other properties the implementation deems to serve an alias role
+        
         :param curie:
         :return:
         """

tests/test_implementations/test_pronto.py

@@ -100,6 +100,14 @@ def test_synonyms(self):
                                     'cellular component',
                                     'cell or subcellular entity',
                                     'subcellular entity'])
+        syns = self.oi.aliases_by_curie(NUCLEUS)
+        logging.info(syns)
+        self.assertCountEqual(syns, ['nucleus', 'cell nucleus', 'horsetail nucleus'])
+        syn_pairs = list(self.oi.alias_map_by_curie(NUCLEUS).items())
+        self.assertCountEqual(syn_pairs,
+                              [('oio:hasExactSynonym', ['cell nucleus']),
+                               ('oio:hasNarrowSynonym', ['horsetail nucleus']),
+                               ('rdfs:label', ['nucleus'])])
 
     def test_subsets(self):
         oi = self.oi

tests/test_implementations/test_sparql.py

@@ -57,6 +57,15 @@ def test_synonyms(self):
         syns = self.oi.aliases_by_curie(CELLULAR_COMPONENT)
         logging.info(syns)
         assert 'cellular component' in syns
+        assert 'cellular_component' in syns
+        syns = self.oi.aliases_by_curie(NUCLEUS)
+        logging.info(syns)
+        self.assertCountEqual(syns, ['nucleus', 'cell nucleus', 'horsetail nucleus'])
+        syn_pairs = list(self.oi.alias_map_by_curie(NUCLEUS).items())
+        self.assertCountEqual(syn_pairs,
+                              [('oio:hasExactSynonym', ['cell nucleus']),
+                               ('oio:hasNarrowSynonym', ['horsetail nucleus']),
+                               ('rdfs:label', ['nucleus'])])
 
     def test_all_entity_curies(self):
         curies = list(self.oi.all_entity_curies())

tests/test_implementations/test_ubergraph.py

@@ -46,6 +46,17 @@ def test_synonyms(self):
         logging.info(syns)
         assert 'cellular component' in syns
 
+    @unittest.skip('This test is too rigid as synonyms are liable to change')
+    def test_synonyms_granular(self):
+        syns = self.oi.aliases_by_curie(NUCLEUS)
+        logging.info(syns)
+        self.assertCountEqual(syns, ['nucleus', 'cell nucleus', 'horsetail nucleus'])
+        syn_pairs = list(self.oi.alias_map_by_curie(NUCLEUS).items())
+        self.assertCountEqual(syn_pairs,
+                              [('oio:hasExactSynonym', ['cell nucleus']),
+                               ('oio:hasNarrowSynonym', ['horsetail nucleus']),
+                               ('rdfs:label', ['nucleus'])])
+
     def test_definition(self):
         defn = self.oi.get_definition_by_curie('GO:0005575')
         logging.info(defn)