texflatten

#!/bin/sh -f
#
# script for recursive flattening of tex/latex/bibtex into a single
# file (written to stdout)
#
# Usage: texflatten.sh <infile.tex>
#
# Important: for citations to work, there must be an up-to-date
# <infile.bbl>. I.e.: run bibtex first on the input files before
# flattening.


# bib file is main file basename .bbl
bibfile="${1%.*}.bbl"

# main gawk script
read -r -d '' awkflatten <<EOF
function insertFile(filename) {
    print "%%%%%%%%%%%%%%%%%%%%%%%%%%%%%"
    print "%%% BEGIN "filename
    print "%%%%%%%%%%%%%%%%%%%%%%%%%%%%%"
    while ((getline line < filename) > 0)
        if( !match( line,/^%/ ) )
            print line
    close(filename)
    print "%%%%%%%%%%%%%%%%%%%%%%%%%%%%%"
    print "%%% END "filename
    print "%%%%%%%%%%%%%%%%%%%%%%%%%%%%%"
}

{
    if( match( \$0,/^[^%]*\\\\(include|input)\{([^\}]+)/,includes ) ) {
        filename= includes[2]".tex"
        insertFile(filename)
    } else if( match( \$0,/^[^%]*\\\\bibliography\{/ ) ) {
        insertFile("$bibfile")
    } else if( !match( \$0,/^%/ ) ) {
        # strip full line comments while we are at it
        print
    }
}
EOF

cd `dirname $1`

# iterative flattening until there are no more include, input and
# bibliography statements left (except those commented out)
infile=`mktemp flatten_input.XXXXXXX`
outfile=`mktemp flatten_output.XXXXXXX`
cp `basename $1` $outfile
while [ "$(egrep '^[^%]*\\(include|input|bibliography){' $outfile)" != "" ]; do
    (>&2 echo "flattening iteration...")
    cp $outfile $infile
    gawk "$awkflatten" $infile > $outfile
done

# output result
echo "%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%"
echo "%%% File generated by texflatten"
echo "%%% (you probably want to edit the source files instead)"
echo "%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%"
cat $outfile

# clean up temp files
rm $infile $outfile