Add arXiv hyperlink to bibtex in under 1 seconds

Add hyperlinks of arXiv numbers to all the bibtex entries in a bib file in under 1 second. This is accomplished by a script.  The user only needs to have a bibtex file which already has the eprint entry for every bibtex item. The rest would be handled by the script.

This procedure is not necessary for some packages and bst files. it becomes necessary when specialized citation packages, like natbib, are used. Natbib generally do not include the hyperlink to the eprint number and some bst files for natbib like pjgsm.bst would completely ignore the eprint field.

The following script would amend this situation. It basically hard codes the hyperlinked arXiv number  into the journal entry. So hen the bibtex is run the journal entry would include the hyperlinked arXiv number. I wrote this script with bibtex from the NASA ADS server in mind. Bibtex from other sources might need some modification.

Critical comments and suggested improvements are welcome.

NOTE !!! I deliberately deleted 2 crucial but trivial lines of code as a puzzle to the readers. !!!

If anyone can figure it out and post it in the comments or if nobody figure it out and there are multiple request in the comments section  then I would include them back.

!!! Enjoy !!!

==============================================



#/bin/bash
# Usage : ./eprint.sh <filename>
# v1.2 by JOnes 
# This script takes the arXiv number from the eprint entry in the bibtex and hard codes a hyperlinked arXiv number to  the journal entry of the bibtex. This is done for every bibtex entry. It has a failsafe and will not act upon a file on which it is already run; running twice on the same file would not have a detrimental effect on the output.


if [[ "$#" -ne 1 ]]; then
  echo  "Usage : ./eprint.sh <filename> 
# v1.2 by JOnes 
# This script takes the arXiv number from the eprint entry in the bibtex and hard codes a hyperlinked arXiv number to  the journal entry of the bibtex. This is done for every bibtex entry."
  exit
fi


isbib=`grep -i "@article" $1` 


if [ ! -e $1  ] ; then
  echo "File do not exist."
  exit
elif [[ ! $isbib ]] ; then
  echo "This file cannot be worked upon."
  exit
else
  cp $1 .tmp
fi


epntlist=`cat .tmp | grep "eprint" | cut -f2 -d"=" | sed 's/\"//g' | sed 's/\,//g' | sed 's/{//g' | sed 's/}//g'`


if [[ ! $epntlist  ]] ; then
 exit
fi




cp .tmp .tmp2 


for ii in $epntlist
do
kk1=`cat .tmp | grep $ii -A2 -B4| grep journal | grep -v "href" `


kk2=`echo $kk1 | sed "s|}\",|; \\href{http://arxiv.org/abs/$ii}{$ii} }\",|"`


if [[ `echo $kk2 | grep  href` ]] ; then
  sed "s|$kk1|$kk2|" .tmp > .tmp2
fi




kk2=`echo $kk1 | sed "s|\",|; \\href{http://arxiv.org/abs/$ii}{$ii} \",|"`


if [[ `echo $kk2 | grep  href` ]] ; then
  sed "s|$kk1|$kk2|" .tmp > .tmp2
fi


kk2=`echo $kk1 | sed "s|},|; \\href{http://arxiv.org/abs/$ii}{$ii} },|"`


if [[ `echo $kk2 | grep  href` ]] ; then
  sed "s|$kk1|$kk2|" .tmp > .tmp2
fi


# 's/dsas/afdsa' would not work as --- 1)bash treats anything inside ' ' single quotes as literal string. when " " double quotes are used then only do bash replace the values of the variables inside the strings 2) \ has to be escaped by 3 backslashes \\\ . 3) Since \href uses the \ character once it cannot be used as a delimiter of sed arguments hence | | is used.
#ll=`echo $kk | grep href`


done


cat .tmp | sed 's/ href/ \\href/g' > .tmp2
mv .tmp2 converted.bib


echo "done.. See file converted.bib."



==============================================