pre_tokenizer_whitespace function

This pre-tokenizer simply splits using the following regex: \w+|[^\w\s]+