Check for a valid string on the regex transformer. If not a valid string,...

Check for a valid string on the regex transformer.  If not a valid string, then the value is ignored and the original record returned. (#136)

databuilder/transformer/regex_str_replace_transformer.py

@@ -18,6 +18,8 @@ class RegexStrReplaceTransformer(Transformer):
     """
     Generic string replacement transformer using REGEX.
     User can pass list of tuples where tuple contains regex and replacement pair.
+
+    Any non-string values will be ignored.
     """
     def init(self, conf):
         # type: (ConfigTree) -> None
@@ -27,7 +29,12 @@ class RegexStrReplaceTransformer(Transformer):
     def transform(self, record):
         # type: (Any) -> Any
         val = getattr(record, self._attribute_name)
-        if six.PY2 and isinstance(val, six.text_type):
+
+        if val is None or not isinstance(val, six.string_types):
+            return record
+
+        # Encode unicode string
+        if six.PY2:
             val = val.encode('utf-8', 'ignore')
 
         for regex_replace_tuple in self._regex_replace_tuples:

setup.py

@@ -2,7 +2,7 @@ import os
 from setuptools import setup, find_packages
 
 
-__version__ = '1.4.3'
+__version__ = '1.4.4'
 
 
 requirements_path = os.path.join(os.path.dirname(os.path.realpath(__file__)), 'requirements.txt')

tests/unit/transformer/test_regex_str_replace_transformer.py

@@ -9,6 +9,33 @@ class TestRegexReplacement(unittest.TestCase):
 
     def test(self):
         # type: () -> None
+        transformer = self._default_test_transformer()
+
+        foo = Foo('abc')
+        actual = transformer.transform(foo)
+
+        self.assertEqual('bba', actual.val)
+
+    def test_numeric_val(self):
+        # type: () -> None
+        transformer = self._default_test_transformer()
+
+        foo = Foo(6)
+        actual = transformer.transform(foo)
+
+        self.assertEqual(6, actual.val)
+
+    def test_none_val(self):
+        # type: () -> None
+        transformer = self._default_test_transformer()
+
+        foo = Foo(None)
+        actual = transformer.transform(foo)
+
+        self.assertEqual(None, actual.val)
+
+    def _default_test_transformer(self):
+        # type: () -> RegexStrReplaceTransformer
         config = ConfigFactory.from_dict({
             'regex_replace_tuple_list': [('a', 'b'), ('c', 'a')],
             'attribute_name': 'val'
@@ -17,10 +44,7 @@ class TestRegexReplacement(unittest.TestCase):
         transformer = RegexStrReplaceTransformer()
         transformer.init(config)
 
-        foo = Foo('abc')
-        actual = transformer.transform(foo)
-
-        self.assertEqual('bba', actual.val)
+        return transformer
 
 
 class Foo(object):