Add function for the n-gram spam detection heuristic workflow
add distinct version of ParseR count_ngram which excludes 'optional & non-optional', so that spam regex and text variable line up - some stop words and other things like punctuation are removed for better matching of patterns