Ugly code to get rough CSV into wrapped EAD XML chunks

gistfile1.rb

#!/usr/bin/env ruby
# encoding: utf-8

require 'csv'
require 'pp'
require 'active_support/core_ext'
require 'nokogiri'


f = nil; File.open('/home/jhbrown/Dropbox/kaye.csv') do |file|
  f = CSV.parse(file.read.encode('UTF-8', :invalid => :replace))
  # , :headers => true, :header_converters => :symbol)
end

f.each_with_index do |line,index|
  line.each_with_index.map do |line_el, line_el_index|
    if line_el.blank? && index != 1
      f[index][line_el_index] = f[index - 1][line_el_index]
    end
  end
  
  line[(0..3)].map(&:to_s).join.match(/\((?<folders>\d+)[^\)]*?\)/).tap do |o|
    if o.nil?
      f[index] << ""
    else
      f[index] << "#{o[:folders]} folder" if o[:folders] == "1"
      f[index] << "#{o[:folders]} folders" if o[:folders] != "1"
    end 
  end
  
end

csv_string = CSV.generate(:encoding => 'UTF-8') do |csv|
  f.each do |v|
    csv << v
  end
end


puts csv_string

# f = CSV.read('/home/jhbrown/projects/kaye_parsed.txt')
f = CSV.parse(csv_string)
f.map!{|x| x.map{|y| y.nil? ? y : y.strip } }
f.map!{|x| x.map{|y| y == "none" ? "" : y } }

xdoc = Nokogiri::XML::Builder.new(:encoding => 'UTF-8') do |xml|
  xml.root {
    f.each_with_index do |line,index|
      unless line[2].nil? || index == 0
        xml.c02(:level => 'item') {
          xml.did {
            unless line[0..1].reject(&:blank?).empty?
              xml.scopecontent(:encodinganalog => "subject") {
                xml.list {
                  xml.item {
                    xml.text line[0..1].reject(&:blank?).join("--")
                  }
                }
              }
            end        
            xml.unitdate(:encodinganalog => "date") {
              xml.text line[2]
            }
            xml.unittitle(:encodinganalog => "title") {
              xml.text(line[3].blank? ? "General" : line[3])
            }
            xml.container(:type => "Box") { 
              xml.text line[4]
            }
            unless line[5].blank?
              xml.physdesc(:encodinganalog => "format") {
                xml.text line[5]
              }
            end      
          }
        } 
      end
    end
  }
end

puts xdoc.to_xml