Pattern application

[zverok]

After the rules parsing is done (#3) and text parsing is done (#4), there should be a Pattern class developed, able to find rule patterns in input text.

The patterns seem to consist of tokens, each of them could be:

• literal word;
• regexp;
• POS tagged word;
• exception in any of above forms.

Though, there probably can be other, more complicated forms. Example (negative lookbehind): <token>all<exception scope="previous" regexp="yes">in|of|not|at</exception></token> (I assume, it reads as "all word, unless in|of|not|at words were before it").

The result of this task should be:

• clarification of the form in which parsed input text should be stored (so Pattern could be effectively applied);
• Pattern able to search through text and return some PatternMatch object (with reference to matched text and numerized matched tokens).