script for extracting apa; I might write a post at some point about how it all works. it's mainly a personal script; it's very raw; use at own risk

bibmash

#!/bin/bash      
                                            
bibtexstylefile=apalike
bibtexdatabase=$BIB
pdfdirectory=$BIBDIR
pdfprogram=evince
pandocoutputformat=plain
# returntask
# bibtex : return bibtex record (default)
# format : return 
# key    : return citation keys
# pdf    : open pdfs
returntask=bibtex

# summary files
extractedbibtexfile=$(tempfile).bib
extractedkeyfile=$(tempfile)
extractedformatfile=$(tempfile)
extractedpdfnamesfile=$(tempfile)

help()
{
   cat <<END
Usage: bibmash [options] search-terms

-h, --help              get this help
-p, --pdf               open pdfs
-f, --format <format>   format reference output using pandoc 
                        e.g., plain, markdown, html, latex
-t,                     format references using pandoc in plain text
                        same as -f plain
-k, --keys              return bibtex keys
-e <search-term>        search terms
-s, --select            select one pdf to open, citation to return, or bibtex to
                        return from list
-b <bibtex-database>    specificy bibtex database
                        defaults to script configured setting
END
    exit
}

# OPTS=$(getopt -n bibmash -o f:k:e:b:hpt --long help,format,keys,select "$@")
# if [ $? != 0 ] ; then echo "Terminating..." >&2 ; exit 1 ; fi
# eval set -- "$OPTS"

for i in "$@"
do
	case "$i" in
		-p|--pdf) returntask=pdf; shift ;;
        -e) searchterm="$2"; shift 2 ;;
		-f|--format) pandocoutputformat="$2"; returntask=format; shift 2 ;;
		-k|--keys) returntask=key; shift ;;
        -t) returntask=format; pandocoutputformat=plain; shift ;;
	 	-h|--help) help ;; 
	esac
done

# cat $extractedbibtexfile
# extract subset of bibtex
bibtool '--select{"'"$searchterm"'"}' $bibtexdatabase > $extractedbibtexfile
bibtool $extractedbibtexfile | grep '^@' | awk '{print $2}' | \
    sed 's/,//' | tr '\n' ' ' > $extractedkeyfile
bibtex2html -nokeys -o - -s $bibtexstylefile -nodoc -q  $extractedbibtexfile  | \
    sed '/file was generated/,$d; 1,9d' | \
    pandoc -r html -w $pandocoutputformat > $extractedformatfile

# Format extracted bibtex using bibtex style with html markup
# Use sed to remove supplementary html output provided by bibtex2html 
if [ $returntask = bibtex ]; then
    cat $extractedbibtexfile
    exit
fi

if [ $returntask = format ]; then
    cat $extractedformatfile
    exit
fi


if [ $returntask = pdf ]; then
    for i in $(cat $extractedkeyfile)
    do
        find $pdfdirectory -iname $i.pdf >> $extractedpdfnamesfile
    done


    # if only one result is returned, then open it
    #if [ $(echo results | wc -w) -eq 1 ]
    #then
        #$pdfprogram $results
        #exit
    #fi


    # open selected matches in browser
    
    select i in $(cat $extractedpdfnamesfile | tr '\n' ' ' )
    do 
        $pdfprogram $i 2> /dev/null &
    done
    exit
fi

if [ $returntask = key ]; then
    cat $extractedkeyfile
    exit
fi
# Creating a local bibtex file based on aux file
# bibtool -x document.aux -o local.bib

# Extracting a single bibtex reference based on a part or all of a bibtex key
# bibtool -X smith $BIB

# ideas for converting article output to plain text or markdown format
# bibmash warning|pandoc -f html -w plain
# bibmash warning|pandoc -f html -w markdown


# TODO
# * Extract just keys
# * Extract keys in \cite command
# * Open pdf
# modify output formatting: remove pdf note and abstract