Fix serialization bug (OpenNMT#1188)

* In dataset base, remove __reduce_ex__ and override __getattr__.
``torchtext.Dataset.__getattr__`` is a generator. That doesn't
play well with pickle. Returning a generator (when appropriate)
seems to fix the issue without changing API.

onmt/inputters/dataset_base.py

@@ -50,16 +50,6 @@ class DatasetBase(Dataset):
         the same structure as in the fields argument passed to the constructor.
     """
 
-    def __getstate__(self):
-        return self.__dict__
-
-    def __setstate__(self, _d):
-        self.__dict__.update(_d)
-
-    def __reduce_ex__(self, proto):
-        # This is a hack. Something is broken with torch pickle.
-        return super(DatasetBase, self).__reduce_ex__()
-
     def __init__(self, fields, src_examples_iter, tgt_examples_iter,
                  filter_pred=None):
 
@@ -90,6 +80,15 @@ def __init__(self, fields, src_examples_iter, tgt_examples_iter,
 
         super(DatasetBase, self).__init__(examples, fields, filter_pred)
 
+    def __getattr__(self, attr):
+        # avoid infinite recursion when fields isn't defined
+        if 'fields' not in vars(self):
+            raise AttributeError
+        if attr in self.fields:
+            return (getattr(x, attr) for x in self.examples)
+        else:
+            raise AttributeError
+
     def save(self, path, remove_fields=True):
         if remove_fields:
             self.fields = []