As a user I would like a mode that more aggressively predicts author strings #242

Given the following assumptions are always true:

• A string contains one or more of name, author, and year
• That string contains only those elements (i.e. there is never going to be some element that is not one of those three)

Then I would like to set an option to more agressively predict the author element. False positives are not a major concern, the benefit of good match will be seen downstream via a edit-distance calculation element.

The canonical example is: s = aus Jones.

When I do ::Biodiversity::Parser.parse(s).dig(:authorship, :normalized) then I get parsed: false. I would like to get "Jones".