A lot of published papers contain 2 seperate columns of text, I've encountered issues sometimes that Weava doesn't recognise these two columns and treats them as one big chunk of text, making it impossible to highlight a single sentence without taking some random sentence from another column with it.
Edit: This is not Weava's mistake, it was due to the format of that paper. If I opened it in other readers it also didn't recognise the columns...
Thank you for helping Weava improve, and we agree that there ought to be more and better solutions to cater to unique PDFs that may have unique alignments or ways of portraying their text.
While, as you said, this is a PDF-specific problem per se, we will still keep the ticket up as we may investigate further in the future on how we may overcome such challenges to give you a better service.