Merge pull request #1021 from simonwoerpel/fix/levensthein-segfault

Fix levensthein segmentation fault in invalid name picking
alephdata · Feb 27, 2023 · 58ba901 · 58ba901
2 parents 741972d + 788458a
commit 58ba901
Show file tree

Hide file tree

Showing 2 changed files with 12 additions and 3 deletions.
diff --git a/followthemoney/types/name.py b/followthemoney/types/name.py
@@ -1,8 +1,8 @@
-from typing import Dict, List, Optional, Sequence, TYPE_CHECKING, Union
-from banal import first
+from typing import TYPE_CHECKING, Dict, List, Optional, Sequence, Union
+
+from Levenshtein import distance, setmedian
 from normality import slugify
 from normality.cleaning import collapse_spaces, strip_quotes
-from Levenshtein import distance, setmedian
 
 from followthemoney.types.common import PropertyType
 from followthemoney.util import dampen
@@ -57,6 +57,9 @@ def pick(self, values: Sequence[str]) -> Optional[str]:
             lookup.setdefault(norm, [])
             lookup[norm].append(value)
 
+        if not normalised:
+            return None
+
         norm = setmedian(normalised)
         if norm is None:
             return None

diff --git a/tests/types/test_names.py b/tests/types/test_names.py
@@ -22,6 +22,12 @@ def test_pick(self):
         values = ["Robert Smith", "Rob Smith", "Robert SMITH"]
         self.assertEqual(names.pick(values), "Robert SMITH")
 
+        # handle dirty edgecases
+        values = ["", "(", "Peter"]
+        self.assertEqual(names.pick(values), "Peter")
+        values = ["", "("]
+        self.assertEqual(names.pick(values), None)
+
     def test_domain_validity(self):
         self.assertTrue(names.validate("huhu"))
         self.assertFalse(names.validate(""))