Ruby script to process LaTeX files for conversion to .docx using Pandoc as a transport.

pandoc.rb

# Usage:
# => ruby pandoc.rb -i <file_name>(.tex) (options for pandoc)
# => minimum options for pandoc: --bibliography=<bibliography_file.bib> --csl=<mla or chicago>.csl

# Installed by running 'gem install bibtex' (may need to append sudo)
require 'bibtex'

default_bibliography = "./transhumanism.bib" # change this to whatever

if ARGV.find_index("-i")
	file = ARGV[ARGV.find_index("-i") + 1]
	options = {}
	ARGV.delete("-i")
	ARGV.delete(file)
else
	raise "No Input File Provided"
	exit
end

if options.has_key? "bibliography"
    bibliography = BibTeX.open(options["bibliography"])
else
    bibliography = BibTeX.open(default_bibliography)
end

if file =~ /\.tex$/
	options["file_name_base"] = file.sub(/\.tex$/,"")
else
	options["file_name_base"] = file
	file = "#{file}.tex"
end

if !File.exists? file
	raise "File #{file} Not Found"
	exit
end

pandoc_options = ARGV.join(" ")

new_latex_content = ""
labels = {}
$chapter = 1
figure = 1
section = 1

# Recursively scan in a file for labels
def scan_in_file(file_name, indent)
	if File.exists? "./#{file_name}.tex"
		include_file = File.read("./#{file_name}.tex")
		figure = 1
		section = 1
		
        # Pull in other included files
        include_file.gsub!(/^(\t*)\\input\{([^}]+)\}/) do |match|
            scan_in_file($2, $1)
        end
        
		# Collect label information:
		include_file.gsub!(/\t*\\([^{]*)\{(.+)\}\n\s*\\label\{(.*)\}/) do |match|
			$labels[$3] = {}
			if $1 == "caption"
				$labels[$3]["number"] = "#{$chapter}.#{figure}"
				figure += 1
			elsif $1 == "section"
				$labels[$3]["number"] = "#{$chapter}.#{section}"
				section += 1
			else
				$labels[$3]["number"] = "#{$chapter}"
			end
			
			$labels[$3]["name"] = $2
			match.sub("\\label{#{$3}}","")
		end
        
		line = "\n#{indent}#{include_file}".gsub(/\n/,"\n#{indent}")
    else
        puts "Could not find #{file_name}.tex"
        line = ""
	end
    line
end

File.open("#{options["file_name_base"]}.tex", "r") do |latex|
    # Detect if the file we're processing is not in the same directory as the parser
    file_base_dir = (File.dirname("#{options["file_name_base"]}.tex") == ".") ? ("") : File.dirname("#{options["file_name_base"]}.tex") + "/"
	while line = latex.gets	
		if line =~ /^(\t*)\\ChapterInput\{([^}]+)\}/ or line =~ /^(\t*)\\input\{([^}]+)\}/
			line = scan_in_file(file_base_dir + $2, $1)
            $chapter += 1
		end
		new_latex_content += line
	end
end

# Correctly insert label text
new_latex_content.gsub!(/\\(name){0,1}ref\{([^}]+)\}/) do |match|
	if $1 == "name" and !$labels[$2].nil?
		$labels[$2]["name"]
    elsif !$labels[$2].nil?
		$labels[$2]["number"]
    else
        "LABEL NOT FOUND"
	end
end

#puts new_latex_content

# enumerate* and itemize* make pandoc freak out.
new_latex_content.gsub!(/(enumerate|itemize)\*/,"\\1")

# Look up titles
new_latex_content.gsub!(/\\citetitle\{([^}]+)\}/) do |match|
    key = $1
    if bibliography[key].type == :book
        "\\emph{#{bibliography[key].title}}"
    else
        "``#{bibliography[key].title}''"
    end
end

# Fix wonkey citation format for pandoc-citeproc
new_latex_content.gsub!(/(citep|autocite|cite)\[([0-9, ]+)\]\{(\S+)\}/) do |match|
	command = $1
	pages = $2
	cite_key = $3
	if pages =~ /\,/
		page_numbers = pages.split(/, */)
		pages = page_numbers.map { |page| "p. #{page}" }.join(", ")
	else
		pages = "p. #{pages}"
	end
	"#{command}[#{pages}]{#{cite_key}}"
end

# Provide absolute file names for figures:
new_latex_content.gsub!(/(\t+)\\includegraphics(.*)\{([^}]+)\}/) do |match|
	file_name = $3
	["png", "eps", "pdf", "jpg"].each do |extension|
		if File.exists? "#{file_name}.#{extension}"
			file_name = "#{file_name}.#{extension}"
		end
	end
	"#{$1}\\includegraphics#{$2}{#{file_name}}"
end

# Fix epigraphs:
new_latex_content.gsub!(/(\t+)\\epigraph\{(.+)\}\{(.+)\}/, "\\1\\begin{center}\n\\1\t\\emph{\\2 \\linebreak[4] --- \\3}\n\\1\\end{center}")

File.open("#{options["file_name_base"]}-pandoc.tex", "w") do |pandoc|
	pandoc.puts new_latex_content
end

exec "pandoc -s --smart #{options["file_name_base"]}-pandoc.tex -o #{options["file_name_base"]}.docx #{pandoc_options}"
#exec "pandoc -s -f latex -t markdown+citations #{options["file_name_base"]}-pandoc.tex -o #{options["file_name_base"]}.md #{pandoc_options}"
rm "#{options["file_name_base"]}-pandoc.tex"