limpiar_df flexibility & functionality

[jpcompartir]

Decide what to do with numbers, punctuation etc.

Problem being that removing them isn't always clearcut and most likely shouldn't be included in limpiar_df, or if they are to be, should be the non-default options i.e. user should make a conscious choice that they want to remove numbers and punctuation. Functions should mimic textmining (tm) package, but without the whitespace issues.