pdf2doi

#!/usr/bin/env sh
#: Extract the DOI from a PDF.
#:
#: Usage:
#:
#:    pdf2doi [filename]
#:
#: If no filename is specified, read from stdin.
#:
#: Requirements: pdfinfo (install poppler_utils), pdftotext (install xpdf or python's pdftotext).

method1() {
  pdfinfo $1 | grep '^Subject: .*;10\.' | head -n 1 | sed 's/.*;10\./10./'
}

method2() {
  pdfinfo -meta $1 | grep 'prism:doi' | head -n 1 | sed 's_.*<prism:doi>__; s_</prism:doi>.*__'
}

method3() {
  pdftotext -l 3 $1 - | grep -e '10\.[0-9]*/' | head -n 1 | sed 's/.*\(10\.[0-9]*\)/\1/'
}

DOI=$(method1 $1)

if [ -z "${DOI}" ]; then
    # echo "Trying method 2 for $1"
    DOI=$(method2 $1)
fi

if [ -z "${DOI}" ]; then
    # echo "Trying method 3 for $1"
    DOI=$(method3 $1)
fi

if [ -z "${DOI}" ]; then
    DOI=???
fi

echo "DOI for $1 is ${DOI}"