Request: Better font selection in PDFs

Summary:

When using the internal PDF importer, it is frequently the case that fonts in PDFs are given strange names. In my experience, the "closet-named" font option fails to select the right font most of the time. While of course the Poppler import works really well, the text is essentially uneditable, resulting in the default sans-serif font being used.

It would be really nice if the internal importer had a way of automatically selecting the best font. I am thinking that test text could be rendered from all system fonts, and some sort of overlap fraction could be calculated with the Poppler output.