Bug in calculating all_by_all_pairwise_similarity (#800)

* Updated semsimian version

* poetry lock --no-update

* corrected score assignment

* formatted

pyproject.toml

@@ -23,7 +23,7 @@ kgcl-schema = "^0.6.9"
 
 funowl = ">=0.2.0"
 gilda = {version = ">=1.0.0", optional = true}
-semsimian = {version = ">=0.2.16", optional = true}
+semsimian = {version = ">=0.2.18", optional = true}
 kgcl-rdflib = "0.5.0"
 llm = "^0.14"
 html2text = {version = "*", optional = true}

src/oaklib/implementations/semsimian/semsimian_implementation.py

@@ -215,6 +215,7 @@ def all_by_all_pairwise_similarity(
             minimum_resnik_threshold=min_ancestor_information_content,
             # predicates=set(predicates) if predicates else None,
         )
+
         logging.info("Post-processing results from semsimian")
         for term1_key, values in all_results.items():
             for term2_key, result in values.items():
@@ -228,15 +229,15 @@ def all_by_all_pairwise_similarity(
                             iter(ancestor_set)
                         ),  # TODO: Change this: gets first element of the set
                     )
-                    sim.jaccard_similarity = jaccard
-                    sim.ancestor_information_content = resnik
-                    sim.phenodigm_score = phenodigm_score
+
                 else:
                     sim = TermPairwiseSimilarity(
                         subject_id=term1_key, object_id=term2_key, ancestor_id=OWL_THING
                     )
-                    sim.jaccard_similarity = 0
-                    sim.ancestor_information_content = 0
+                sim.jaccard_similarity = jaccard if jaccard is not None else 0.0
+                sim.ancestor_information_content = resnik if resnik is not None else 0.0
+                sim.phenodigm_score = phenodigm_score if phenodigm_score is not None else 0.0
+
                 yield sim
 
     def termset_pairwise_similarity(

tests/test_cli.py

@@ -1522,7 +1522,9 @@ def test_annotate_file(self):
             ],
         )
         print("STDERR", result.stdout)
-        err = "\n".join([line for line in result.stderr.split("\n") if not line.startswith("WARNING")])
+        err = "\n".join(
+            [line for line in result.stderr.split("\n") if not line.startswith("WARNING")]
+        )
         self.assertEqual("", err)
         self.assertEqual(0, result.exit_code)
         with open(outfile) as stream: