speed up and fixes

textattack/attack_recipes/alzantot_2018.py

@@ -60,4 +60,4 @@ def Alzantot2018(model):
     #
     search_method = GeneticAlgorithm(pop_size=60, max_iters=20)
 
-    return Attack(goal_function, constraint, transformation, search_method)
+    return Attack(goal_function, constraints, transformation, search_method)

textattack/constraints/modification_constraint.py

@@ -13,7 +13,7 @@ class ModificationConstraint(Constraint):
     def __call__(self, x, transformation):
         """ Returns the word indices in x which are able to be modified """
         if not self.check_compatibility(transformation):
-            return True
+            return set(range(len(x.words)))
         return self._get_modifiable_indices(x)
 
     def _get_modifiable_indices(x):

textattack/search_methods/beam_search.py

@@ -23,7 +23,7 @@ class BeamSearch(SearchMethod):
             potential_next_beam = []
             for text in beam:
                 transformations = self.get_transformations(
-                        text, original_text=original_tokenized_text)
+                        text, original_text=initial_result.tokenized_text)
                 for next_text in transformations:
                     potential_next_beam.append(next_text)
             if len(potential_next_beam) == 0:

textattack/search_methods/genetic_algorithm.py

@@ -135,7 +135,7 @@ class GeneticAlgorithm(SearchMethod):
 
     def __call__(self, initial_result):
         self.original_tokenized_text = initial_result.tokenized_text
-        self.correct_output = intial_result.output
+        self.correct_output = initial_result.output
         neighbors_len = self._get_neighbors_len(self.original_tokenized_text)
         pop = self._generate_population(neighbors_len)
         cur_score = initial_result.score

textattack/shared/attack.py

@@ -76,7 +76,7 @@ class Attack:
         """
         if not self.transformation:
             raise RuntimeError('Cannot call `get_transformations` without a transformation.')
-        
+       
         transformations = np.array(self.transformation(text, 
                                    modification_constraints=self.modification_constraints, **kwargs))
         if apply_constraints:
@@ -138,7 +138,7 @@ class Attack:
         if final_result.succeeded:
             return SuccessfulAttackResult(initial_result, final_result)
         else:
-            return FailedAttackResult(intial_result, final_result)
+            return FailedAttackResult(initial_result, final_result)
  
     def _get_examples_from_dataset(self, dataset, num_examples=None, shuffle=False,
             attack_n=False, attack_skippable_examples=False):

textattack/shared/tokenized_text.py

@@ -162,7 +162,7 @@ class TokenizedText:
         """
         final_sentence = ''
         text = self.text
-        new_attack_attrs = deepcopy(self.attack_attrs)
+        new_attack_attrs = dict()
         new_attack_attrs['stopword_indices'] = set()
         new_attack_attrs['modified_indices'] = set()
         new_attack_attrs['newly_modified_indices'] = set()

textattack/transformations/transformation.py

@@ -14,8 +14,7 @@ class Transformation:
         else:
             indices_to_modify = set(indices_to_modify)
         for constraint in modification_constraints:
-            if constraint.check_compatibility(self):
-                indices_to_modify = indices_to_modify & constraint(tokenized_text, self)
+            indices_to_modify = indices_to_modify & constraint(tokenized_text, self)
         transformed_texts = self._get_transformations(tokenized_text, indices_to_modify)
         for text in transformed_texts:
             text.attack_attrs['last_transformation'] = self

textattack/transformations/word_swap_gradient_based.py

@@ -67,6 +67,7 @@ class WordSwapGradientBased(Transformation):
         # grad differences between all flips and original word (eq. 1 from paper)
         vocab_size = lookup_table.size(0)
         diffs = torch.zeros(len(indices_to_replace), vocab_size)
+        indices_to_replace = list(indices_to_replace)
         for j, word_idx in enumerate(indices_to_replace):
             # Get the grad w.r.t the one-hot index of the word.
             b_grads = emb_grad[word_idx].view(1,-1).mm(lookup_table_transpose).squeeze()