Handle the case of invalid slot names in cardinality_scope.

Explicitly look for invalid slot names in the `scope` list given to the
`infer_cardinality` method, and forcibly remove them.

The presence of invalid slot names does _not_ change the result of
cardinality computations, but explicitly looking for such slots allows
us to do two things:

(1) warn users that they're using invalid slot names;
(2) avoid writing invalid slot names in the `cardinality_scope` slot.

Author: gouttegd

src/sssom/util.py

@@ -422,6 +422,12 @@ def infer_cardinality(self, scope: List[str] = None) -> None:
         subjects_by_object: dict[str, set[str]] = {}  # Unique subjects for any given object
         objects_by_subject: dict[str, set[str]] = {}  # Unique objects for any given subject
 
+        schema = SSSOMSchemaView()
+        unknown_slots = [slot for slot in scope if slot not in schema.mapping_slots]
+        if len(unknown_slots) > 0:
+            logging.warning(f"Ignoring invalid slot name(s): {unknown_slots}.")
+            scope = list(set(scope) - set(unknown_slots))
+
         # Helper function to transform a row into a string that represents
         # a subject (or object) in a given scope; `side` is either `subject`
         # or `object`.

tests/test_utils.py

@@ -12,6 +12,7 @@
 from sssom_schema import slots as SSSOM_Slots
 
 from sssom.constants import (
+    CARDINALITY_SCOPE,
     CREATOR_ID,
     MAPPING_CARDINALITY,
     OBJECT_ID,
@@ -623,3 +624,13 @@ def test_infer_scoped_cardinality(self) -> None:
         msdf.infer_cardinality(["object_source"])
         expected = ["1:1", "1:1", "1:1", "1:1", "1:1", "1:1"]
         self.assertEqual(expected, list(msdf.df[MAPPING_CARDINALITY].values))
+
+        msdf.infer_cardinality(["object_source", "not_a_valid_slot_name"])
+        # should yield the same result as above
+        self.assertEqual(expected, list(msdf.df[MAPPING_CARDINALITY].values))
+
+        msdf.infer_cardinality(["not_a_valid_slot_name"])
+        # should be equivalent to an empty scope
+        expected = ["1:n", "1:n", "1:n", "1:n", "1:n", "1:n"]
+        self.assertEqual(expected, list(msdf.df[MAPPING_CARDINALITY].values))
+        self.assertNotIn(CARDINALITY_SCOPE, msdf.df.columns)