Parse a Word OpenXML document file with a predefined structure

parse-word-structure.rb

#!/usr/bin/env ruby

require 'nokogiri'
require 'zip/zipfilesystem'

file = ARGV[0] or abort("Usage: #$PROGRAM_NAME <file.docx>")

Zip::ZipFile.open(file) do |zip|

  doc  = Nokogiri::XML(zip.file.read('word/document.xml'))
  rels = Nokogiri::XML(zip.file.read('word/_rels/document.xml.rels'))

  doc.xpath('//w:tr').each do |row|
    fields = row.xpath('.//w:tc').map do |cell|
      [].tap do |ret|
        if link = cell.xpath('.//w:hyperlink').first
          id = link['id']
          ret << rels.css("Relationship[Id=#{id}]").first['Target']
        end

        ret << cell.xpath('.//w:t').text
      end
    end

    puts fields.join(',')
  end

end