Skip to content

Latest commit

 

History

History
98 lines (79 loc) · 4.67 KB

epub-recompression.md

File metadata and controls

98 lines (79 loc) · 4.67 KB
Raw

EPUB Recompression

Trimming needlessly large ebooks

Sometimes you'll download an EPUB file and it's just way too big. Often this can be fixed relatively easily, as you'll see here. Just follow these steps:

  1. Copy the offending EPUB into a temporary directory that you are comfortable with using.

  2. unzip humongous.epub -d huge, making the obvious substitutions. I use a single lowercase word from the title for the temporary directory I unpack the contents into. And yes, really. EPUB files are glorified ZIP archives.

  3. Find out where the images are stored and cd into the directory containing them. (I have only ever come across cases where a single directory holds all images.) The exact location varies, but finding it should be easy as you will see the paths scrolling by as the contents of the EPUB are extracted. Once you find it, You might want to run du -ch in the directory to get a sense of how large the images are at this point. Typically, they are JPEGs. If they are PNGs, skip to step 6; otherwise proceed.

  4. In my experience, huge EPUB files with JPEG images are a consequence of a needlessly high compression setting. Everyone hates JPEGs that look like they've had dirt smeared on them, but it usually really is possible to reduce the compression without any appreciable loss of quality. So what you want to do will look like this:

    for img in *.jpg; do convert -quality 70 $img "reduced-$img"; done &

    This isn't a recipe set in stone. The quality can be changed up or down to fit your needs, but I find that 70 is a sensible default. Also, if the extension .jpeg is being used instead of .jpg, make the obvious change. When all is said and done, there will be a set of images in the folder with names identical to those you started with but for the prefix reduced-.

  5. The reduced images will only take up more space in the archive if they don't overwrite the ones that are already there. They can replace the original images in one fell swoop, like so, remembering to do this after the last step runs to completion:

    for img in reduced-*; do mv $img $(echo $img | sed 's/^reduced-//'); done

    Now go to step 8.

  6. If there are PNG images in the image folder, something I have only come across once, then—at least in the sole instance I came across—converting them all to JPEGs will result in massive saving of space with no real loss of quality. In the image folder:

    for img in *.png; do convert -quality 70 $img $(basename $img .png).jpg; done &

    As before, quality may be adjusted to suit your purposes. After the background processes run to completion, rm *.png to get rid of the PNG images.

  7. With JPEG recompression, the original filenames will be saved. Not so with conversion from PNG. Find the directory that contains the text of the EPUB. It will have a large number of files with the extension .html or similar (such as .xhtml). As far as I know, as with images, the text of the EPUB is all stored in one directory. cd into this directory, then run the following, making any necessary change appropriate for the file suffix used in this directory:

    for page in *.html; do ruby -p -i -e 'gsub(/<img(.*?)src="(.*?)\.png"/, "<img\\1src=\"\\2\.jpg\"")' $page; done

    I believe this regex should be fully appropriate. On the one hand, it will prevent instances of img tags being discussed in the text from being modified, because, in those cases, the angle bracket must be escaped as &lt;. On the other hand, it is robust enough to account for an alt attribute or the like intervening between img and the src attribute.

  8. If you followed my advice in step 3, cd to the image directory if you are not already in it and run du -ch again to get a sense of how much space you'll save when the EPUB is recompressed. Of course this figure won't be exact because it doesn't take compression into account, but it will give you an idea of how well you did.

  9. Now cd to the directory that you extracted the EPUB into, in other words, the directory that followed the -d option to unzip, and zip -r recompressed.epub *, which will pack everything into a new EPUB file. Obviously, the name should be altered to fit or changed later. You now have a fully functioning EPUB!

  10. Give your new EPUB the once-over. Is the new EPUB substantially smaller than the old one? Do the images display properly? Is the image quality adequate? If so, congratulations on your new, smaller EPUB! If not, try again with different parameters. If you suspect something is wrong with my procedure, please inform me and I will see what can be done about it.