Currently we save PDF text as File metadata. Consequently the text is invisible to the advanced search page. See https://forum.omeka.org/t/pdf-text-not-being-indexed-for-search/3080
We should consider adding an option to select where to save the PDF text: as File metadata or as Item metadata. If so, we'll need to devise a way to handle existing text when users switch between the two.