ans_epub_fix

#!/usr/bin/env ruby
# frozen_string_literal: true

# HELIO-3901. Script to repair ANS EPUBs.

require 'optparse'
require 'ostruct'
require 'os'

# Process the script parameters.
options = OpenStruct.new
option_parser = OptionParser.new do |opts|
  opts.banner = "Usage: #{File.basename(__FILE__)} <epub_file> [<epub_file>...]"
  opts.on_tail('-h', '--help', 'Print this help message') do
    puts opts
    exit 0
  end
end
option_parser.parse!(ARGV)
if ARGV.count < 1
  puts option_parser.help
  return
end

# Process the command line parameters.
epub_file_list = ARGV

require 'logger'
script_logger = Logger.new(STDOUT)
script_logger.formatter = proc do |severity, datetime, progname, msg|
  "#{severity}: #{msg}\n"
end

# Determine the root directory of the code base.
script_dir = File.expand_path(File.dirname(__FILE__))
root_dir = File.dirname(script_dir)

require_relative File.join(root_dir, 'lib', 'epub')
require_relative File.join(root_dir, 'lib', 'remediation')

processors = {
          default: UMPTG::Remediation::DefaultProcessor.new
      }

epub_file_list.each do |epub_file|
  epub_file = File.expand_path(epub_file)
  unless File.file?(epub_file)
    script_logger.fatal("Invalid EPUB \"#{epub_file}\".")
    next
  end
  script_logger.info("Using EPUB \"#{epub_file}\".")

  epub = UMPTG::EPUB::Archive.new(epub_file: epub_file)

  # Process the epub. Returned is a hash table where each
  # item key is an EPUB entry name and the item value is
  # a list of processing actions.
  action_map = UMPTG::EPUB::Processor.process(
        epub: epub,
        entry_processors: processors,
        pass_xml_doc: true
      )
  action_map.each do |entry_name, proc_map|
    script_logger.info("#{entry_name}")

    xml_doc = proc_map[:xml_doc]
    update_entry = false
    proc_map.each do |key,action_list|
      next if key == :xml_doc

      action_list.each do |action|
        case action.status
        when UMPTG::Action.COMPLETED
          update_entry = true
          script_logger.info(action.to_s)
        else
          script_logger.warn(action.to_s)
        end
      end
    end

    # Update the entry in the EPUB. Replace old entry with
    # the new one.
    if update_entry
      epub.add(entry_name: entry_name, entry_content: UMPTG::XML.doc_to_xml(xml_doc))
    end
  end

  # Fix toc.ncx
  toc_entry = epub.entry("OEBPS/toc.ncx")
  toc_doc = UMPTG::XML.parse(xml_content: toc_entry.content)
  n_list = toc_doc.xpath("//*[local-name()='navPoint']/*[local-name()='content']")

  update_entry = false
  n_list.each do |n|
    src = n['src'].strip
    mdata = src.match(/([^#]*)#(.*)/)
    unless mdata.nil?
      if mdata[2].to_i > 0
        newSrc = "#{mdata[1]}\##{File.basename(mdata[1],'.*')}_#{mdata[2]}"
        script_logger.info("#{src}==>#{newSrc}")
        n['src'] = newSrc
        update_entry = true
      end
    end
  end
  if update_entry
    epub.add(entry_name: toc_entry.name, entry_content: UMPTG::XML.doc_to_xml(toc_doc))
  end

  # Fix toc.xhtml
  toc_entry = epub.entry("OEBPS/toc.xhtml")
  toc_doc = UMPTG::XML.parse(xml_content: toc_entry.content)
  n_list = toc_doc.xpath("//*[local-name()='a']")

  update_entry = false
  n_list.each do |n|
    href = n['href'].strip
    mdata = href.match(/([^#]*)#(.*)/)
    unless mdata.nil?
      if mdata[2].to_i > 0
        newHref = "#{mdata[1]}\##{File.basename(mdata[1],'.*')}_#{mdata[2]}"
        script_logger.info("#{href}==>#{newHref}")
        n['href'] = newHref
        update_entry = true
      end
    end
  end
  if update_entry
    epub.add(entry_name: toc_entry.name, entry_content: UMPTG::XML.doc_to_xml(toc_doc))
  end

  fixed_epub_file = File.join(File.dirname(epub_file), File.basename(epub_file, ".*") + "_fixed.epub")
  script_logger.info("Saving #{fixed_epub_file}")
  epub.save(
        epub_file: fixed_epub_file
      )
end