Script for converting PDF books to readable format on small screens.

pdf_convert_for_e_books.rb

# Testes on ruby-2.4.0
# With rmagick 2.16.0
# And ImageMagick 6.9.7-4
require 'rmagick'
require 'fileutils'

INPUT_FILE='/Users/kzhidovinov/Downloads/code_complete.pdf'
OUTPUT_FILE = '/Users/kzhidovinov/Downloads/code_complete_1.pdf'

TMP_DIR = File.join(File.dirname(__FILE__), 'temp')
# Size of page
TARGET_HEIGHT = 560
TARGET_WIDTH = 760
DENSITY = 300

FileUtils.mkpath(TMP_DIR)
system("rm #{File.join(TMP_DIR, '*')}")

output_mask = File.join(TMP_DIR, 'p%04d.png')
# Use ghostscript util to split PDF into PNG images.
system("gs -dNOPAUSE -dBATCH -sDEVICE=pngalpha -sOutputFile=\"#{output_mask}\" -r#{DENSITY} #{INPUT_FILE}")

# Will contain page file names
out_filenames = []

# Process all pages
Dir.glob("#{TMP_DIR}/p*.png").each_with_index do |file, idx|

  puts "Processing #{file}"

  # read image from file
  image = Magick::Image.read(file).first

  image.density = "#{DENSITY}x#{DENSITY}"
  # trim spaces
  image.trim!(true)
  # make image gray
  image.colorspace = Magick::GRAYColorspace
  # resize image to width = TARGET_WIDTH
  image.change_geometry!("#{TARGET_WIDTH}x#{TARGET_HEIGHT}^") do |w, h, img|
    img.resize!(w, h)
  end

  # Split current pages to fit target page size.
  height = image.rows
  i = 0
  while i * TARGET_HEIGHT < height
    filename = File.join(TMP_DIR, "#{'%06d' % (idx*100+i)}.out.pdf")

    out_filenames << filename

    out_img = image.crop(Magick::NorthGravity, 0, [i * TARGET_HEIGHT - 10, 0].max, TARGET_WIDTH, TARGET_HEIGHT, true)
                  .extent(TARGET_WIDTH, TARGET_HEIGHT)
                  .rotate(90)
                  .sharpen(5, 10.0)

    out_img.define('pdf:use-trimbox', true)
    out_img.density = "#{DENSITY}x#{DENSITY}"
    out_img.write(filename)

    i = i + 1
  end

end

# merge pages into PDF
system("gs -dNOPAUSE -dBATCH -sDEVICE=pdfwrite -sOutputFile=\"#{OUTPUT_FILE}\" #{out_filenames.join(' ')}")