A ruby command line script to generate jpg's from the first pages of pdf files.

convert_pdf.rb

#!/usr/bin/env ruby

# This require that you have ImageMagick and Ghostscript installed and available on your PATH
# It also assumes you have you pdf files in `pdfs` folder
# Note: Since it generates jpg's it does not handle transparency very well
#       modify the .jpg to .png to generate png files.

Dir['pdfs/*.pdf'].each do |pdf_file|
  jpg_file = pdf_file.chomp(File.extname(pdf_file)) + ".jpg"
  system("convert -density 144 #{pdf_file}[0] -resize 200x300 #{jpg_file} ")
end

I prefer GraphicsMagick, the ImageMagick fork with better code quality, more sane command line options, and a less annoying developer. :)

$ gm convert -quality 85 -thumbnail x600 -flatten 64661.pdf\[0\] cover.jpg

The results are very similar obviously! One tip, I prefer the "x600" notation, which gives 600px height at whatever width is appropriate for the aspect ratio—hey, these thumbnails will be around for decades, on god-only-knows-what-resolution devices right?

edit: Also, there are diminishing returns in JPEG quality after around 85 or so, so I always use 82–85.

These thumbnails will be around for decades, on god-only-knows-what-resolution devices right?

Usually, keeping a high-resolution copy that you could manipulate on the fly is a good idea. Progress in the Responsive Images community is promising in getting rid of the whole device-screen nightmare.

We could always re-process them, you're right. For now, they need to be high-quality enough to be used wherever our partners might use them, for example:

https://ccafs.cgiar.org/publications/how-countries-plan-address-agricultural-adaptation-and-mitigation

The size of these is more or less ok, but the quality is horrible.