extractOne is sensitive to ordering

[rafikmatta]

extractOne changes result based on ordering, and my guess is this is due to Levenstein distance being equal. Maybe a new distance measure is necessary?

e.g.
process.extractOne("C++",["C#","C++","C",".NET"])
Out - ('C#', 100)

process.extractOne("C++",["C++","C#","C",".NET"])
Out - ('C++', 100)

Maybe a heuristic needs to come in for exact match?