I tested part of this on Windows in a git bash shell:

$ pdftotext -v
pdftotext version 0.14.5
Copyright 2005-2010 The Poppler Developers - http://poppler.freedesktop.org
Copyright 1996-2004 Glyph & Cog, LLC

and then this works:
$ pdftotext.exe kitchin-2015-examp.pdf - | grep DOI
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.5b00538 • Publicati
on Date (Web): 11 May 2015
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is
an optional service offered
sharing site which assigned the data set a DOI. 19 An alternative data repositor
y could be an
institutional data repository which also provides a DOI for citing. It remains t
o be seen if
stores that provide a citable DOI for the data set. The point is that this appro
ach is very

However, I see that in emacs, it appears another pdftotext is getting used, which does not work.
~/Desktop $ pdftotext -v
pdftotext version 2.03
Copyright 1996-2003 Glyph & Cog, LLC

that is the same version that runs in a cmd shell.

The version that seems to work for me is at "C:\Program Files (x86)\Git\bin\pdftotext".

I added a variable to set the location of this program, and then rely on the executable path. It was a little tricky to set this right, note the escaped quotes.

#+BEGIN_SRC emacs-lisp
(setq pdftotext-executable "\"C:/Program Files (x86)/Git/bin/pdftotext\"")
#+END_SRC


It turned out there was another issue with the uri in the drag-n-drop needing to be unescaped on windows because of the : in the path, which is also done now. 

I pushed these out. hopefully that makes some progress for you! or at least gives some hints on where to look for the problem.

Merry Xmas!

John

-----------------------------------
Professor John Kitchin 
Doherty Hall A207F
Department of Chemical Engineering
Carnegie Mellon University
Pittsburgh, PA 15213
412-268-7803

On Wed, Dec 23, 2015 at 6:56 PM, Andreas Kiermeier <andreas.kiermeier@gmail.com> wrote:
Hi John,
similar doi problem here.
My configuration is:
  • Windows 10
  • GNU Emacs 24.5.1 (i686-pc-mingw32) of 2015-04-11 on LEG570
  • Miktex (up-to-date) with pdftotext v0.32..0
  • org-ref v 0.5.0 (from melpa)
Running pdftotext from the command line works and produces a text file with doi info.
Let me know what I can do to help to test things.
Thanks.
Andreas
PS: Thank you for org-ref ... an awesome Xmas present!


On 24 December 2015 at 08:00, John Kitchin <jkitchin@andrew.cmu.edu> wrote:
Thanks! Do you know if you have pdftotext working on your machine? The pdf drag-n-drop works by converting the pdf to text, and than matching a pattern to find a doi. If none is found, you get the message you noted. The url dnd works similarly, but there are a bunch of recipes for what to match depending on the base of the url.

I did that on a Mac, and I haven't tested it on a windows or Linux machine. 

John

-----------------------------------
Professor John Kitchin 
Doherty Hall A207F
Department of Chemical Engineering
Carnegie Mellon University
Pittsburgh, PA 15213
412-268-7803

On Wed, Dec 23, 2015 at 4:22 PM, marvin doyley <marvinpas@gmail.com> wrote:
Hi John,

I am playing with org-ref,  the melpa version.

Everything works fine except, one thing.  When I drag a pdf to an empty bibtex file it doesn’t extract the doi ( no doi found in the file:///). I tried it pdf you used  in your video (Examples of Effective Data Sharing in Scientific Publishing), but I got the same error. Dragging the url to the bibtex file also doesn’t works for me.

cheers,
M

PS by the way, your video was excellent :)