feat(medcat): CU-869aknekf Add mapping to ontologies (#147)

* CU-869aknekf: Remove simple_hash config option.

It's an unused relic from v1

* CU-869aknekf: Add config option for mapping to other ontologies

* CU-869aknekf: Implement mapping to other ontologies

* CU-869aknekf: Fix typing issue with ontology mappings

* CU-869aknekf: Add a few tests to check mappings are appropriately added

* CU-869aknekf: Remove extra whitespace

* CU-869aknekf: Removed TODO that has been addressed regarding output

* CU-869aknekf: Move logged message from debug to warning

Author: mart-r

medcat-v2/medcat/cat.py

@@ -523,22 +523,34 @@ def _get_entity(self, ent: MutableEntity,
             'context_similarity': ent.context_similarity,
             'start': ent.base.start_char_index,
             'end': ent.base.end_char_index,
-            # TODO: add additional info (i.e mappings)
-            # for addl in addl_info:
-            #     tmp = self.cdb.addl_info.get(addl, {}).get(cui, [])
-            #     out_ent[addl.split("2")[-1]] = list(tmp) if type(tmp) is
-            # set else tmp
             'id': ent.id,
-            # TODO: add met annotations
-            # if hasattr(ent._, 'meta_anns') and ent._.meta_anns:
-            #     out_ent['meta_anns'] = ent._.meta_anns
             'meta_anns': {},
             'context_left': left_context,
             'context_center': center_context,
             'context_right': right_context,
         }
         # addons:
         out_dict.update(self.get_addon_output(ent))  # type: ignore
+        # other ontologies
+        if self.config.general.map_to_other_ontologies:
+            for ont in self.config.general.map_to_other_ontologies:
+                if ont in out_dict:
+                    logger.warning(
+                        "Trying to map to ontology '%s', but it already "
+                        "exists in the out dict, so unable to add it. "
+                        "If this is for an actual ontology that shares a "
+                        "name with something else, cosider renaming the "
+                        "mapping in `cdb.addl_info`")
+                    continue
+                addl_info_name = f"cui2{ont}"
+                if addl_info_name not in self.cdb.addl_info:
+                    logger.warning(
+                        "Trying to map to ontology '%s' but it is not set in "
+                        "addl_info so unable to do so", ont)
+                    continue
+                ont_map = self.cdb.addl_info[addl_info_name]
+                ont_values = ont_map.get(cui, [])
+                out_dict[ont] = ont_values  # type: ignore
         return out_dict
 
     def get_addon_output(self, ent: MutableEntity) -> dict[str, dict]:

medcat-v2/medcat/config/config.py

@@ -252,11 +252,17 @@ class General(SerialisableBaseModel):
     map_cui_to_group: bool = False
     """If the cdb.addl_info['cui2group'] is provided and this option enabled,
     each CUI will be mapped to the group"""
-    simple_hash: bool = False
-    """Whether to use a simple hash.
+    map_to_other_ontologies: list[str] = ["opcs4", "icd10"]
+    """Which other ontologies to map to if possible.
 
-    NOTE: While using a simple hash is faster at save time, it is less
-    reliable due to not taking into account all the details of the changes."""
+    This will force medcat to include mapping for other ontologies in
+    its outputs. It will use the mappings in `cdb.addl_info["cui2<ont>"]`
+    are present.
+
+    NB!
+    This will only work if the `cdb.addl_info["cui2<ont>"]` exists.
+    Otherwise, no mapping will be done.
+    """
 
     class Config:
         extra = 'allow'

medcat-v2/tests/test_cat.py

@@ -159,6 +159,39 @@ def test_entities_in_correct_order(self):
                 cur_start = ent.base.start_char_index
 
 
+class InferenceIntoOntologyTests(TrainedModelTests):
+    ont_name = "FAKE_ONT"
+
+    @classmethod
+    def setUpClass(cls):
+        super().setUpClass()
+        # create mapping
+        cls.ont_map = {
+            cui: [f"{cls.ont_name}:{cui}"]
+            for cui in cls.model.cdb.cui2info
+        }
+        # add to addl_info
+        cls.model.cdb.addl_info[f"cui2{cls.ont_name}"] = cls.ont_map
+        # ask to be mapped
+        cls.model.config.general.map_to_other_ontologies.append(cls.ont_name)
+
+    def assert_has_mapping(self, ent: dict):
+        # has value
+        self.assertIn(self.ont_name, ent)
+        val = ent[self.ont_name]
+        # 1 value
+        self.assertEqual(len(val), 1)
+        # value in our map
+        self.assertIn(val, self.ont_map.values())
+
+    def test_gets_mappings(self):
+        ents = self.model.get_entities(
+            ConvertedFunctionalityTests.TEXT)['entities']
+        for nr, ent in enumerate(ents.values()):
+            with self.subTest(f"{nr}"):
+                self.assert_has_mapping(ent)
+
+
 class CATIncludingTests(unittest.TestCase):
     TOKENIZING_PROVIDER = 'regex'
     EXPECT_TRAIN = {}