Handling of non-ASCII / unicode characters

[Michael-E-Rose]

Sometimes pubmed_parser returns text as is, sometimes it uses unidecode().

First, I think it should be consistent throughout the package.
Second, I'm not sure unidecode() is still needed or even desired - recipients should be allowed to work with the original text.

What do you think, @titipata ?