Refactor the SQLizer code to be more generic

[Wes-Reid]

Right now the SQLizer project knows too much. It shouldn't care about where the file is that it's working with or where to put the scripts or even be concerned with actually making the files. I think it should only worry about turning an excel spreadsheet into SQL statements.

It should probably just take the file as a stream and return a stream or streams of SQL. Let the caller handle getting and creating the files. That would also make it easier to move the project into a web app (which is another to do).