Enhance the way we specify Persian words

[amirhoseinsalimi]

In Persian, we have two separators: space and half-space. Currently, we have an issue in words-not-to-follow and words-to-follow: We have to specify words in both forms, once with the common space character and once with a half-space. People use these words in their tweets - interchangeably - so the bot has to be sensitive on both of them, but polluting the aforementioned files with redundant data is not a good idea (which is the current implementation). So it's good to write a new function (or modify this function) to do this job. In brief, this new implementation should:

• Convert any half-space to a normal space (And then we can only enter words w/ half-spaces and the bot takes care of the rest)
• Convert dashes into underscore (bonus)