MTG Event Parser

Pdf turned out to be unstable (tables can wrap over several pages), hence I decided to implement it with html instead. I also refactored your code quite a bit to make it more reusable.

Also I just wrote a test for the parsing itself. For the other parts I think it makes sense to write more reusable tests, since I assume all Parsers will either use links or files to import the data. So we can have just 2 generic tests for that.

Maybe you can have a look at the parser_names/parser_views issue. Basically I want to make adding further parsers as easy as possible, by defining the names and views of parsers in a single place. However since I need access to the names during the creation of the views I run into a circular import dependency. Hence I had to split up parser_names and parser_views into 2 lists which looks ugly.