Add less exhaustive modes of searching pattern using regular expressions

Currently regexp version of find pattern algorithm performs an exhaustive search trying to detect any possible occurence of the string matching the regular expression. This may lead to situations when a lot of redundant job is done. For instance when user inputs the "N+" pattern and the reference sequence contains a substring consisting only of "N" symbols and having length ~1000 then the find pattern algorithm yields ~10000 result annotations on this region. It seems more reasonable to add new modes of the regexp search that will produce results in succession trying to match the largest and the smallest possible substrings.