Word bounaries for unicode chars

[sr9yar]

\b are not working for languages other than english. Which makes it impossible to use at all. Simply removing them seems to work fine though.

[waseem18]

\b are not working for languages other than english

@sr9yar I'm not sure if node-rake works for languages other than English and even if it works I'm not sure about it's effectiveness.

I'm speculating if removing \b is going to affect the performance of keyword generation.