added support for stratified x-validation, if the input is as an orange table or dbcontext

workflows/models.py

@@ -231,18 +231,19 @@ class Workflow(models.Model):
         current_iteration = 0
 
         # create folds
-        rand.seed(input_seed)
-        rand.shuffle(input_list)
 
         folds = []
         if hasattr(input_list, "get_items_ref"):
+            import orange 
             # Orange table on input, so we cannot do slices
-            indices = range(len(input_list))
-            folds_indices = [indices[i::input_fold] for i in range(input_fold)]
-            folds = []
-            for fold_indices in folds_indices:
-                folds.append(input_list.get_items_ref(fold_indices))
+            indices = orange.MakeRandomIndicesCV(input_list, randseed=input_seed)
+            for i in range(input_fold):
+                output_train = input_list.select(indices, i, negate=1)
+                output_test = input_list.select(indices, i)
+                folds.append((output_train, output_test))
         else:
+            rand.seed(input_seed)
+            rand.shuffle(input_list)
             folds = [input_list[i::input_fold] for i in range(input_fold)]
 
         # pass forward the seed
@@ -258,8 +259,13 @@ class Workflow(models.Model):
                     i.save()
 
         for i in range(len(folds)):
-            output_train = folds[:i] + folds[i+1:]
-            output_test = folds[i]
+            #import pdb; pdb.set_trace()
+            if hasattr(input_list, "get_items_ref"):
+                output_test = folds[i][0]
+                output_train = folds[i][1]
+            else:
+                output_train = folds[:i] + folds[i+1:]
+                output_test = folds[i]
             if input_type == 'DBContext':
                 output_train_obj = context.copy()
                 output_train_obj.orng_tables[context.target_table] = output_train