Script that converts biblatex (bibtex) styled latex document bibliography to standard bibitem bibliography (thebibliography)

bibtex-to-latex-bibitem.sh

#!/bin/bash

###############################################################
#                                                             #
# BibTeX (BibLaTeX) to LaTeX \bibitem converter               #
#                                                             #
###############################################################
#                                                             #
# This script converts a file containing                      #
# BibTeX (BibLaTeX) entries to a list of \bibitems,           #
# suitable for use in LaTeX.                                  #
#                                                             #
# Special thanks to this blog post:                           #
# http://timothyandrewbarber.blogspot.com/2011 ->             #
#    /08/fundamental-thinking-convert-bibtex.html             #
# for pointing me in the correct direction                    #
# and giving the instruction on how to do                     #
# this.                                                       #
#                                                             #
# Dependencies: perl, awk, sed, tr, latex, bibtex, bash       #
#                                                             #
# Usage:                                                      #
# ./bibtex-to-latex-bibitem.sh file1.tex file2.tex > out.bib  #
#                                                             #
###############################################################

# config
documentclass="russian"
bibliographystyle="plain"
# it can be something more interesting e.g.:
# bibliographystyle="ugost2008"

# variables
tex_filenames=$@
tmp_dir=$(mktemp -d)
tmp_bib=$tmp_dir/refs.bib
tmp_tex=$tmp_dir/dummy.tex
tmp_tex_basename=$(basename $tmp_tex .tex)

# extract keys
# split keys separated by commas
# uniq without sorting to preserve order
keys=$(cat $tex_filenames | perl -lne 'print join "\n", m/\\cite\{([^}]+)}/g' | sed 's/,\s*/\n/g' | awk '!x[$0]++')

# DEBUG: print found keys
# for key in ${keys[@]}
# do
#     echo $key 1>&2
# done 

echo "Count Keys: " $( expr $(echo "${keys[@]}" | wc -l) - 1 ) 1>&2

# extract bib entries by keys
if [ -f $tmp_bib ]; then
    echo "rm $tmp_bib"
    rm $tmp_bib
fi

# bib_files=("${@:2}")
# Extract name of bib files from tex files
bib_files=$(cat $tex_filenames | perl -lne '/\\bibliography\{([^}]+)}/ or next; print $1' | sed 's/,\s*/\n/g' | sed '/\.bib$/! s/$/\.bib/' | tr '\n' ' '| awk '!x[$0]++')

for key in $keys
do
    sed -n "/${key}/,/^\s*}\s*$/p" ${bib_files} >> $tmp_bib
done

# Create template
printf "\\\\documentclass[${documentclass}]{article}\\n" > $tmp_tex 
printf '\\usepackage[utf8]{inputenc}\n' >> $tmp_tex
printf '\\usepackage{babel}\n' >> $tmp_tex
printf '\\begin{document}\n' >> $tmp_tex
printf '\\nocite{*}\n' >> $tmp_tex
printf "\\\\bibliography{${tmp_bib}}\\n" >> $tmp_tex
printf "\\\\bibliographystyle{${bibliographystyle}}\\n" >> $tmp_tex
printf '\\end{document}\n' >> $tmp_tex

# Create references
cd $tmp_dir
latex  ${tmp_tex_basename}  &>/dev/null
bibtex ${tmp_tex_basename}  &>/dev/null
bibtex ${tmp_tex_basename}  &>/dev/null
latex  ${tmp_tex_basename}  &>/dev/null

cat ${tmp_tex_basename}.bbl

See also https://tex.stackexchange.com/questions/154215/biblatex-biber-to-bibitem

There is an alternative to parsing tex-file for \cites and extracting necessary items by keys from extracted bib-files. And it is:

biber --output_format=bibtex tex-file.bcf

Dummy file:

\documentclass{article}
\begin{document}
\nocite{*}
\bibliography{main_biber.bib}
\bibliographystyle{unsrt}
\end{document}

So v2 of this script should do it with biber instead of straightforward tex-file parser.