Automatic creation of article item from a pdf

marcbernot
As many other people I guess, I have a lot of pdf files in various (dis)organized directories. It would be really great if you could create an article item from a pdf file, i.e. Zotero would extract automatically authors, title, abstract, extracts ... then find the article on the web (through google scholar it should be ok), check by himself or by asking the user (by displaying the first page of the pdf next to the webpage) that he has done a good job and create the item element with the pdf attached to it.
Of course you could then do the same for a whole directory full of pdf.
Finally what I'd like is the same as what itune is doing when it is collecting and organizing all music files you have on your computer. This would help a lot to move painlessly from the old system to Zotero, before getting the good habit to use it on a daily basis.
  • Matthias
    While the PDF file format offers some rudimentary tags for metadata storage, these are very seldomly (if at all) used by publishers, and there's no way to reliably and consistently extract things like an abstract from a PDF.

    See also previous discussions on this topic, e.g.:

    http://forums.zotero.org/discussion/255/
  • marcbernot
    edited November 16, 2007
    Ok, but I do not propose to extract metadata from the pdf file. I propose to extract some data from the pdf, look for that data on google scholar which then permits to obtain the web address of the article and so reliable metadata.
    Please try by yourself : take 10 or even less consecutive words in the main part of any article and look for that chain in google scholar. It really often boils things down to one response!! I'm sure it is possible to make this quite reliable.
    I often proceed this way to locate an article, it is just a pain to do it one at a time.

This is an old discussion that has not been active in a long time. Instead of commenting here, you should start a new discussion. If you think the content of this discussion is still relevant, you can link to it from your new discussion.