Synonym sync: Keep source case if acronym

- Update: Addressed an issue where we were preferring the Mondo capitalization even though it was incorrect, i.e. cases where the synonym was an acronym and capitalized in the source, but was not capitalized in Mondo.

src/scripts/sync_synonym.py

@@ -50,7 +50,7 @@
     'related_synonym_type': '>AI oboInOwl:hasSynonymType SPLIT=|',
 }
 SORT_COLS = ['case', 'mondo_id', 'source_id', 'synonym_scope_source', 'synonym_type', 'synonym_type_mondo', 'synonym']
-
+MONDO_ABBREV_URI = 'http://purl.obolibrary.org/obo/mondo#ABBREVIATION'
 
 def _query_synonyms(ids: List[CURIE], db: SqlImplementation) -> pd.DataFrame:
     """Get synonym triples from sqlite DB"""
@@ -76,7 +76,7 @@ def _add_synonym_type_inferences(row: pd.Series, exclusions: Set[str]) -> str:
 
     # Acronym: uppercase, no numbers, no whitespace, <10 chars
     if syn not in exclusions and syn.isupper() and not any(char.isspace() for char in syn) and len(syn) < 10:
-        types.add('http://purl.obolibrary.org/obo/mondo#ABBREVIATION')
+        types.add(MONDO_ABBREV_URI)
 
     return '|'.join(types)
 
@@ -131,6 +131,10 @@ def _common_operations(
 
     # Format
     if not df_is_combined:
+        # - Acronyms: Use source case
+        #   This operation prevents capitalization from being lost, as sometimes Mondo has used lowercase.
+        df['synonym'] = df.apply(lambda row:
+            row['synonym_case_source'] if MONDO_ABBREV_URI in row['synonym_type'] else row['synonym'], axis=1)
         # - Add ROBOT columns for each synonym scope
         synonym_scopes = ['exact', 'broad', 'narrow', 'related']
         for scope in synonym_scopes:
@@ -301,7 +305,7 @@ def sync_synonyms(
     source_types_df = source_types_df[source_types_df['synonym_type'] != '']
     # -- property conversions
     source_types_df['synonym_type'] = source_types_df['synonym_type'].apply(lambda x: x.replace(
-        'http://purl.obolibrary.org/obo/OMO_0003012', 'http://purl.obolibrary.org/obo/mondo#ABBREVIATION'))
+        'http://purl.obolibrary.org/obo/OMO_0003012', MONDO_ABBREV_URI))
     # -- merge in synonym types
     source_df = source_df.merge(source_types_df, on=['source_id', 'synonym_scope', 'synonym'], how='left').fillna('')
     # - get synonym_types: inferred