bibtex2markdown.rb

#!/usr/bin/env ruby
# 
# This script is a wrapper around pandoc that uses pandoc's
# builtin citeproc support to generate a markdown bibliography
# from bibtex.
#
# Inspired by Jacob Barney's [bib2mkd][] script.
#
# [bib2mkd]: http://jmbarney.dyndns.org/?/linux/bib2mkd/
#
# Simplest usage:
#
#    cat file.bib | bibtex2markdown
#
# will send to STDOUT a markdown bibliography containing all the
# publications from file.bib on STDOUT.
#
# Filtering:
#
#    cat file.bib | bibtex2markdown smith jones
#
# will send to STDOUT a markdown bibliography containing all the 
# publications from file.bib whose citekeys begin with 
# 'smith' or 'jones'.
#
# Filtering with Regexs:
#
#    cat file.bib | bibtex2markdown ".*nes"
#
# will send to STDOUT a markdown bibliography containing all the
# publications from file.bib whose citekeys begin with something
# that matches the regex ".*nes", e.g., 'jones1992' and 
# 'ernest1878'.
#
# Using a default bibtex file:
#
#    bibtex2markdown -d
#
# If the first argument is '-d', then will use the file given by
# the variable defaultbib, defined below, rather than reading 
# STDIN.

defaultbib="/Users/david/.pandoc/default.bib"

input = ""
bibfile = ""

if ARGV[0] == "-d"
	bibfile = defaultbib
	File.open(defaultbib, 'r') { |bibtex| 
		input = bibtex.read
	}
	ARGV.shift
else
	input = STDIN.read
	tmpbibfile = "/tmp/bibtex2markdown" + Process.pid.to_s + ".bib"
	bibfile = tmpbibfile
	File.open(tmpbibfile, 'w') { |bibtex| 
		bibtex.puts "#{input}"
	}
end

strings=ARGV

if strings.length == 0
	strings = [ "" ]
end

keys = []
strings.each { |match|
    match.chomp
    keys = keys + input.scan(/@.*?\{\s*(#{match}.*?),/i)
}
keys.uniq!
keys.sort!

refs = "@" + keys.join(" @")

bibliography = `echo "#{refs}" | pandoc --bibliography="#{bibfile}" -t markdown --no-wrap 2> /dev/null`

puts bibliography.lines.to_a[2..-1].join