Workflow with the pdfs from arxiv.org and bibtex records

[Igor Shenderovich <shender.i@gmail.com>]

Hi, 

I'm trying to establish a nice workflow to collect all the interesting
papers from arxiv.org and corresponding bibtex records. Arxiv.org is the
server which contains papers in pdf format and short abstract for each
of them (for example, we have abstract here [1] and the corresponding
pdf file here [2]). There is another website, inspirehep.net, which
collects the bibtex records (and much more information) of these
papers (see [3] for the bibtex record for cited paper).

I wrote a python script which downloads the bibtex record and put it in
the file, say, ref.bib. Then it downloads the paper itself (the pdf),
and renames it according to the citation key as in the ref.bib. So
mentioned paper will be renamed as Maldacena:1997re.pdf.

Then I want to setup an org file (say, ref.org), which will contain each
paper as a bibtex record, generally, as org-bibtex is doing, but with
the links to the pdf file. 

So all the desired workflow should look like this:

1) read the paper on the arxiv.org;
2) download the paper and bibtex record using the python script (it can
   be launched by pressing the button);
3) in the same time fire up the emacs window with opened ref.org and
   capture there bibtex record with some link to the pdf file.

So, the question is: how it can be achieved? For the moment I came up
with the additional field in the bibtex record, named "file" with the
link to the pdf file which is understandable by org-mode (so it's
something like this: file = "[[~/ref/Maldacena:1997re.pdf]]"). But how
can I use org-bibtex with the org-capture automatically? 

With best wishes, 
IS

[1] http://arxiv.org/abs/hep-th/9711200
[2] http://arxiv.org/pdf/hep-th/9711200v3.pdf
[3] http://inspirehep.net/record/451647/export/hx