Consider do_lower_case in PreTrainedTokenizer

As pointed out in #1545, when using an uncased model, and adding
a new uncased token, the tokenizer does not correctly identify this
in the case that the input text contains the token in a cased format.

For instance, if we load bert-base-uncased into BertTokenizer, and
then use .add_tokens() to add "cool-token", we get the expected
result for .tokenize('this is a cool-token'). However, we get a
possibly unexpected result for .tokenize('this is a cOOl-Token'),
which in fact mirrors the result for the former from before the new
token was added.

This commit adds
- functionality to PreTrainedTokenizer to handle this
situation in case a tokenizer (currently Bert, DistilBert,
and XLNet) has the do_lower_case=True kwarg by:
    1) lowercasing tokens added with .add_tokens()
    2) lowercasing text at the beginning of .tokenize()
- new common test case for tokenizers

https://github.com/huggingface/transformers/issues/1545

transformers/tokenization_utils.py

@@ -512,6 +512,8 @@ class PreTrainedTokenizer(object):
         to_add_tokens = []
         for token in new_tokens:
             assert isinstance(token, str) or (six.PY2 and isinstance(token, unicode))
+            if self.init_kwargs.get('do_lower_case', False):
+                token = token.lower()
             if token != self.unk_token and \
                     self.convert_tokens_to_ids(token) == self.convert_tokens_to_ids(self.unk_token) and \
                     token not in to_add_tokens:
@@ -605,6 +607,9 @@ class PreTrainedTokenizer(object):
 
             Take care of added tokens.
         """
+        if self.init_kwargs.get('do_lower_case', False):
+            text = text.lower()
+
         def split_on_token(tok, text):
             result = []
             split_text = text.split(tok)