Tokenizing text is something that I tend to do pretty often, typically as the beginning of an NLP workflow. The normal workflow goes something like this: 1. Build a generator to stream in some sentences / documents / whatever. 2. Perform some set of transformations on