How it works:

  • Use find to get a list of all PDF files in the given directory.
  • For each file:
    • Is the file larger than the minimum required file size (5MB)?; skip, if no, continue if yes.
    • Create a low quality variant of the file in /tmp, named „originfilename_low.pdf“
    • Check if low quality variant has a lower file size; if yes move it to the origin folder. If the new file is larger then the origin, remove it.

The script can be found on GitHub, see here.