Open ryanmuller opened 12 years ago
This would basically add functionality to the API to accept an uploaded file - a Skim notes binary. This would be associated with a specific publication, and a specific user. Later, someone who wanted to see "the PDF as that person saw it", would have a local script that could automatically download the PDF from the publishers website (if OA), download the skim notes, and then combine the two in the Skim reader. This could happen from Scrobblr, or from the individual's website (in which case the local script would have to know for which user to ask for the skim files from Scrobblr). On the individual website, the user could even jump straight to a given page in the PDF, based on page numbers.
Does this have to be binary? It'd be nice to store the plain-text, then we could display frequent highlights right on the citation page.
We could easily do both. I think in order to be able to merge it back into a Skim document, it has to be binary.
I have also been thinking about how much information we want to suck in from a users annotations/notes to Scrobblr, and how much to leave, and link to.
From Stian's email
skimdoc.save(skimdoc, {:as => "Skim Notes", :in => "/tmp/skimnotes-binary-pdf"})
exports the skim notes in binary format (just the size of the text + a few kb, not big)
this line
skimdoc.read_notes(skimdoc, {:from => "/tmp/skimnotes-binary-pdf"})
reads the notes, and applies them to the currently open PDF. Works perfectly...