arxiv2pdf

#!/bin/bash

# Script for downloading a PDF from a given arXiv identifier
#      
# Usage:    arxiv2pdf arXiv [filename]
#           arxiv2pdf https://arxiv.org/abs/arXiv [filename]
#      
# If filename is omitted the raw PDF is written to a temporary file.
#      
# Example:  arxiv2pdf 1412.3740
# Example:  arxiv2pdf math/0602029   (old arXiv format)
#
# Notes:


# Issue help message if necessary
if [ $# = 0 ] || [ "x$1" = "x--help" ]; then
	echo "Usage: $(basename $0) arXiv [filename]"
	echo "or:    $(basename $0) https://arxiv.org/abs/arXiv [filename]"
	echo "will try and retrieve the PDF file for the arXiv identifier given."
	echo 
	echo "Example: $(basename $0) 1701.06092"
	echo "         $(basename $0) 1701.06092v1"
	echo "         $(basename $0) https://arxiv.org/abs/1701.06092v1"
	echo "         $(basename $0) https://arxiv.org/abs/1701.06092v1.pdf"
	echo "         $(basename $0) math/0602029"
	echo "         $(basename $0) https://arxiv.org/abs/math/0602029"
	exit 1
fi

# Unify the calling syntax
arxivpattern='([0-9]{4}\.[0-9]{4,}([v][0-9]*)?|[-a-z]{1,}/[0-9]{7})(.pdf)?$'
if [[ $1 =~ $arxivpattern ]]; then
	arxiv=${BASH_REMATCH[1]}
else
	echo "Failed to find arXiv pattern."
	exit 1
fi

# Get web page from arxiv
tmpfile=/tmp/$RANDOM$RANDOM
# echo $1
# echo $arxiv
curl -sL https://arxiv.org/abs/$arxiv > ${tmpfile}

# Get download URL 
downloadpattern="<meta name=\"citation_pdf_url\" content=\"([^\"]+)\""

# Retrieve download URL
# echo $downloadpattern
downloadurl=$(grep --only -E "$downloadpattern" ${tmpfile} | cut -f4 -d\")


# Assign outfile name
if [ -z "$2" ]; then
	# No filename given. Guess one from the arxiv page.
	outfile=$(arxiv2pdfname.sh $arxiv)
	echo ${outfile}
else
	outfile=$2
fi

# Delete designated file, if it exists
rm -f ${outfile}

# Get pdf from arxiv
# echo Executing wget -O $outfile -q --user-agent="Mozilla/5.0" $downloadurl
wget -O ${outfile} -q --user-agent="Mozilla/5.0" ${downloadurl}

echo "File written to \"${outfile}\"."

# Clean up and exit
rm ${tmpfile}
exit 0