Add update-justifcation flag to invert method (#505)

mapping-commons · Mar 22, 2024 · 5db5296 · 5db5296
1 parent f1df5a6
commit 5db5296
Show file tree

Hide file tree

Showing 2 changed files with 21 additions and 5 deletions.
diff --git a/src/sssom/cli.py b/src/sssom/cli.py
@@ -731,12 +731,19 @@ def remove(input: str, output: TextIO, remove_map: str):
     is_flag=True,
     help="If True (default), add inverted mappings to the input mapping set, else, just return inverted mappings as a separate mapping set.",
 )
+@click.option(
+    "--update-justification/--no-update-justification",
+    default=True,
+    is_flag=True,
+    help="If True (default), the justification is updated to 'sempav:MappingInversion', else it is left as it is.",
+)
 @click.option("--inverse-map", help="Path to file that contains the inverse predicate dictionary.")
 def invert(
     input: str,
     output: TextIO,
     subject_prefix: Optional[str],
     merge_inverted: bool,
+    update_justification: bool,
     inverse_map: TextIO,
 ):
     """
@@ -746,6 +753,8 @@ def invert(
     :param subject_prefix: Prefix of all subject_ids.
     :param merge_inverted: If True (default), add inverted dataframe to input else,
                           just return inverted data.
+    :param update_justification: If True (default), the justification is updated to "sempav:MappingInversion",
+                          else it is left as it is.
     :param inverse_map: YAML file providing the inverse mapping for predicates.
     :param output: SSSOM TSV file with columns sorted.
     """
@@ -754,12 +763,15 @@ def invert(
         with open(inverse_map, "r") as im:  # type: ignore
             inverse_dictionary = yaml.safe_load(im)
         inverse_predicate_map = inverse_dictionary["inverse_predicate_map"]
+    else:
+        inverse_predicate_map = None
 
     msdf.df = invert_mappings(
-        msdf.df,
-        subject_prefix,
-        merge_inverted,
-        inverse_predicate_map,
+        df=msdf.df,
+        subject_prefix=subject_prefix,
+        merge_inverted=merge_inverted,
+        update_justification=update_justification,
+        predicate_invert_dictionary=inverse_predicate_map,
     )
     write_table(msdf, output)
 

diff --git a/src/sssom/util.py b/src/sssom/util.py
@@ -1392,6 +1392,7 @@ def invert_mappings(
     df: pd.DataFrame,
     subject_prefix: Optional[str] = None,
     merge_inverted: bool = True,
+    update_justification: bool = True,
     predicate_invert_dictionary: dict = None,
 ) -> pd.DataFrame:
     """Switching subject and objects based on their prefixes and adjusting predicates accordingly.
@@ -1400,6 +1401,8 @@ def invert_mappings(
     :param subject_prefix: Prefix of subjects desired.
     :param merge_inverted: If True (default), add inverted dataframe to input else,
                           just return inverted data.
+    :param update_justification: If True (default), the justification is updated to "sempav:MappingInversion",
+                          else it is left as it is.
     :param predicate_invert_dictionary: YAML file providing the inverse mapping for predicates.
     :return: Pandas dataframe with all subject IDs having the same prefix.
     """
@@ -1458,7 +1461,8 @@ def invert_mappings(
     )
     inverted_df = inverted_df[df.columns]
     inverted_df[PREDICATE_ID] = inverted_df[PREDICATE_ID].map(predicate_invert_map)
-    inverted_df[MAPPING_JUSTIFICATION] = SEMAPV.MappingInversion.value
+    if update_justification:
+        inverted_df[MAPPING_JUSTIFICATION] = SEMAPV.MappingInversion.value
     if not prefixed_subjects_df.empty:
         return_df = pd.concat([prefixed_subjects_df, inverted_df]).drop_duplicates()
     else: