Fix set validator (#52)

* Fix issue with SetValidator and large valid_set (#50)

A new function stringify_set only stringifies n elements of a given set.

* Use itertools.islice for better python2 performance

* Add SetValidator test for coverage

* Fix stringify_set: Return {...} instead of [...]

* Add test cases for stringify_set

Fix stringify_set: Sort the elements before displaying for small sets

* Remove previously introduced test case for set validator

* Add parameter checks for stringify_set

* Rename stringify_set to _stringify_set

* Revert d355def

vladiate/test/test_validators.py

@@ -4,7 +4,7 @@
 from ..validators import (
     CastValidator, EmptyValidator, FloatValidator, Ignore, IntValidator,
     NotEmptyValidator, RangeValidator, RegexValidator, SetValidator,
-    UniqueValidator, Validator
+    UniqueValidator, Validator, _stringify_set
 )
 from ..exceptions import BadValidatorException, ValidationException
 
@@ -216,3 +216,14 @@ def test_base_class_raises():
 def test_all_validators_support_empty_ok(validator_class, args):
     validator = validator_class(*args, empty_ok=True)
     validator.validate('')
+
+
+@pytest.mark.parametrize('a_set, max_len, stringified', [
+    ({'A', 'B', 'C'}, 4, "{'A', 'B', 'C'}"),
+    ({'A', 'B', 'C'}, 3, "{'A', 'B', 'C'}"),
+    ({'A', 'B', 'C'}, 2, "{'A', 'B'} (1 more suppressed)"),
+    ({'A', 'B', 'C'}, 0, "{} (3 more suppressed)"),
+    ({}, 5, "{}"),
+])
+def test_stringify_set(a_set, max_len, stringified):
+    assert _stringify_set(a_set, max_len) == stringified

vladiate/validators.py

@@ -1,4 +1,5 @@
 import re
+from itertools import islice
 
 from vladiate.exceptions import ValidationException, BadValidatorException
 
@@ -70,7 +71,8 @@ def validate(self, field, row={}):
         if field not in self.valid_set:
             self.invalid_set.add(field)
             raise ValidationException(
-                "'{}' is not in {}".format(field, self.valid_set))
+                "'{}' is not in {}".format(field,
+                                           _stringify_set(self.valid_set, 100)))
 
     @property
     def bad(self):
@@ -208,3 +210,21 @@ def validate(self, field, row={}):
     @property
     def bad(self):
         pass
+
+
+def _stringify_set(a_set, max_len, max_sort_size=8192):
+    ''' Stringify `max_len` elements of `a_set` and count the remainings
+
+    Small sets (len(a_set) <= max_sort_size) are displayed sorted.
+    Large sets won't be sorted for performance reasons.
+    This may result in an arbitrary ordering in the returned string.
+    '''
+    # Don't convert `a_set` to a list for performance reasons
+    text = "{{{}}}".format(", ".join(
+        "'{}'".format(value) for value in islice(
+            sorted(a_set) if len(a_set) <= max_sort_size else a_set,
+            max_len)
+    ))
+    if len(a_set) > max_len:
+        text += " ({} more suppressed)".format(len(a_set) - max_len)
+    return text