DdlParsers Parse Improvements

I would like to do 2 things:

1. Add an API method to DdlParsers that would specify which parser to use. Given that a caller knows exactly what parser they want to use, this method will be much more efficient (performance, memory) in that it would bypass parser scoring altogether. This is exactly the scenario for the the new Teiid Designer connection profile DDL import (it wants the Teiid parser to be used).

2. Rewrite existing DdlParsers parse method to be more efficient (performance, memory). One thing that can be done is to replace the inner class ScoredParser with much lighterweight code.

I have local implementations of both changes and have seen good results. The large DDL file I was getting an OutOfMemoryError is now parsing without any issues. And the parse performance has also improved.