[NFA/DFA] Implement transforms for syntactic sugar

[DmitrySoshnikov]

NFA/DFA interpreter supports basic building blocks, such as Kleene-closure: a*. We should transform several constructs which are syntactic sugar to their basic representation, so the interpreter can handle them:

• ~a+ -> aa*~ (done in https://github.com/DmitrySoshnikov/regexp-tree/commit/aa0e12cf3b8e7164c6ed6dd237d7871b81dba246)
• [a-z] -> a|b|c|d ... |z, same for [A-Z], same for [0-9], and other char classes
• support groups in the generator
• etc.

Transforms live in src/interpreter/finite-automaton/transforms.