[Flashwalker]

Helllo from Russia! Nice work . I was used ebook-convert, but it's very slooow and little bugy. Then i wrote a simple bashscript, but script wasn't smart enough without xml parsing. Until i found your script.

Prince is good, nice ability to adapt file for e-read-devices, but needs to remove logo manually (I'm using "Master PDF Editor" http://code-industry.net).

So, i wrote a modification based on your bash script (for my humble printing needs). It uses Fuse-zip (apt-get install fuse-zip) and WkHTMLtoPDF (wk means WebKit http://wkhtmltopdf.org) instead of Unzip and Prince. Fuse-zip mounting a little faster then unzip unpacking and WkHTMLtoPDF have no watermark and it's opensource. But WkHTMLtoPDF have no options for metadata, like in Prince (but i found in your script it doesn't used), and output pdf file size some times a little bigger (may be fonts). To test WkHTMLtoPDF you can try firefox addon "Print pages to PDF", which based on it.

And one sad thing: WkHTMLtoPDF can't print svg cover (cover image in epub usually wrapped in svg tag).
But! I maid a workaround in script (unwrap svg image tag to img tag), and it works, cover prints, but not upscaled to page size.

Here is the script: epub2pdfwk.sh and a little custom style.css.

Dependencies: (in addition to xmlstarlet and optional(for metadata) pdftk and recode):

fuse-zip, wkhtmltopdf, sed.

Options: are same plus 3 additionals:

-n means not load default.css. (Using both -n -S equivalent to use only -s).

-d means don't print cover.

-o accepts wkhtmltopdf options (-o "--grayscale --quiet --page-size A5" Or -o "-gqs A5" Or -o "--page-width 800px --page-height 600px" Or -o "--header-font-name 'Liberation Sans' --footer-font-name 'Liberation Sans'" ...) (By default WkHTMLtoPDF prints to A4 size format).

Already specified WkHTMLtoPDF options can be edited in "WKHTMLTOPDF OPTIONS SETTINGS" section of script in set_wkh2popts() function. They are self explained.

And a little fix and improvement of your epub2pdf.sh: epub2pdf.sh

Cheers!

Just outline how script works:

## Mount epub archive:
	#mkdir "$HOME"/.ZIP
	# umount before mount
	fusermount -uz "$HOME"/.ZIP
	fuse-zip -r "file.epub" "$HOME"/.ZIP -o modules=iconv,from_code=CP866,to_code=UTF8
(iconv here can fix ugly cyrillic filenames in nasty winziped archives)
## Convert html files to pdf
	wkhtmltopdf OEBPS/article0.xhtml ... OEBPS/article100.xhtml wkhtmltopdfoutput.pdf
## or:
	wkhtmltopdf -l --image-dpi 144 --image-quality 90 -T 5 -R 2 -B 5 -L 2 --user-style-sheet "${HOME}/bin/style.css" cover OEBPS/cover.xhtml OEBPS/article0.xhtml ... OEBPS/article100.xhtml ~/Documents/wkhtmltopdfoutput.pdf
## or even:
	wkhtmltopdf --lowquality --image-dpi 144 --image-quality 90 -T 5 -R 2 -B 5 -L 2 --user-style-sheet "${HOME}/bin/style.css" --page-size A5 -O Landscape --load-error-handling ignore --load-media-error-handling ignore --javascript-delay 2500 --header-left '[title]' --header-right '[date]' --header-font-name 'WenQuanYi Micro Hei' --header-font-size 8 --header-line --footer-left '[webpage]' --footer-right '[page] / [topage]' --footer-font-name 'WenQuanYi Micro Hei' --footer-font-size 8 --footer-line --custom-header 'User-Agent' 'Mozilla/5.0 (Windows NT 6.1; rv:25.0) Gecko/20100101 Firefox/25.0' --custom-header-propagation cover OEBPS/cover.xhtml OEBPS/article0.xhtml ... OEBPS/article100.xhtml ~/Documents/wkhtmltopdfoutput.pdf